Path: sparky!uunet!haven.umd.edu!darwin.sura.net!jvnc.net!rutgers!cbmvax!snark!eric
From: e...@snark.thyrsus.com (Eric S. Raymond)
Newsgroups: comp.unix.sysv386,comp.sys.intel,news.answers
Subject: PC-Clone UNIX Hardware Buyer's Guide
Summary: Tips on how and where to buy hardware for your UNIX.
Message-ID: <1h1ycn#8f3N9K0Nn8Lx3DqM2k89cf1F=eric@snark.thyrsus.com>
Date: 6 Jul 92 15:38:45 GMT
Expires: 4 Oct 92 23:00:00 GMT
Sender: e...@snark.thyrsus.com (Eric S. Raymond)
Followup-To: comp.unix.sysv386
Lines: 1128
Approved: news-answers-requ...@MIT.Edu
Archive-name: pc-unix/hardware
Last-update: Mon Jul 5 11:49:45 EDT 1992
Version: 5.0
What's new in this issue:
* Overall reorganization and expansion for standalone distribution.
* More on buying for performance.
* Lots of useful info about power-protecting your system.
* New thinking about power supplies.
* Current trends in the Intel microprocessor market.
* Two competing theories of the EISA advantage.
* A guide to SCSI terminology and options
* Beware the floating-point exception bug!
* The importance of gold-plating. :-)
Gentle Reader: if you end up buying something based on information from this
Guide, please do yourself and the net a favor; make a point of telling the
vendor "Eric's FAQ sent me" or some equivalent. The idea isn't to hype me
personally, I've already got all the notoriety I need from doing things like
_The_New_Hacker's_Dictionary_ --- but if we can show vendors that the Guide
influences a lot of purchasing decisions, I can be a more powerful advocate for
the net's interests, and for you.
0. Contents
I. Introduction
II. Overview of the Market
III. Buying the basics
A. Getting Down to Cases
B. Power Supplies and Fans
C. The Heart Of The Machine
D. Motherboards and BIOSes
E. Peripherals
F. Power Protection
IV. Performance tuning
A. How To Pick Your Processor
B. Of Memory In...
C. Bus Wars
D. IDE vs. SCSI
E. Other Disk Decisions
F. Souping Up X Performance
V. Of Mice and Machines
VI. When, Where and How to Buy
VII. Questions You Should Always Ask Your Vendor
A. Minimum Warranty Provisions
B. Documentation
C. A System Quality Checklist
VIII. Things to Check when Buying Mail-Order
A. Tricks and Traps in Mail-Order Warranties
B. Special Questions to Ask Mail-Order Vendors Before Buying
C. Payment Method
IX. Which Clone Vendors to Talk To
I. Introduction
The purpose of this posting is to give you the background information you need
to be a savvy buyer of 386/486 hardware for running UNIX. It is aimed
especially at hackers and others with the technical skills and confidence to go
to the mail-order channel, but contains plenty of useful advice for people
buying store-front retail. It was formerly part of 386-buyers-faq issues 1.0
through 4.0, and is still best read in conjunction with the pc-unix/software
FAQ descended from that posting.
This document is maintained and periodically updated as a service to the net by
Eric S. Raymond <e...@snark.thyrsus.com>, who began it for the very best
self-interested reason that he was in the market and didn't believe in plonking
down several grand without doing his homework first (no, I don't get paid for
this, though I have had a bunch of free software and hardware dumped on me as a
result of it!). Corrections, updates, and all pertinent information are
welcomed at that address. The editorial `we' reflects the generous
contributions of many savvy USENETters.
This posting is periodically broadcast to the USENET group comp.unix.sysv386
and to a list of vendor addresses. If you are a vendor representative, please
check to make sure any information pertaining your company is current and
correct. If it is not, please email me a correction ASAP. If you are a
hardware-knowledgeable user, please send me any distillation of your
experience that you think might improve this posting.
II. Overview of the Market
The central fact about 386/486 clone hardware that conditions every aspect of
buying it is this: more than anywhere else in the industry, de-facto hardware
standards have created a commodity market with low entry barriers, lots of
competitive pressure, and volume high enough to amortize a *lot* of development
on the cheap.
The result is that this hardware gives you lots of bang-per-buck, and it's
getting both cheaper and better all the time. Furthermore, margins are thin
enough that vendors have to be lean, hungry, and *very* responsive to the
market to survive. You can take advantage of this, but it does mean that much
of the info in the rest of this document will be stale in three months and
completely obsolete in six.
One good general piece of advice is that you should avoid both the highest-end
new-technology systems (those not yet shipping in volume) and the very cheapest
systems put out by vendors competing primarily on price. The problem with
the high end is that it usually carries a hefty "prestige" price premium, and
may be a bit less reliable on average because the technology hasn't been
through a lot of test/improve cycles. The problem with the low end is that
price-cutters sometimes settle for marginal components. UNIX is more
sensitive to hardware flakiness than DOS, so cut-price systems that might
deliver consistently for DOS lemmings can come around and bite you. Use a
little care, and spend the $200-$300 to stay out of the basement. It's worth
it.
The last point deserves a little amplification. In the PC world, there's a lot
of "if it doesn't fail, it's OK". It is common to ignore normal engineering
tolerances --- the allowances for variations in components, temperature,
voltage margins, and the like --- and to assume that anything which doesn't
fail outright must work. Watch out! For example, The ISA bus was originally
designed for 6 MHz. IBM later updated that to 8 MHz, and that's as much of a
standard as there is, yet there are motherboards that will let you (try to!)
run it at 12 MHz --- 50% over spec. Some cards are actually designed to work
at that speed with proper tolerances. Others might work...or they might flake
out when they get warm. Any systems vendor above the fly-by-night level is
going to shoot for a little more reliability than this, burning in systems and
(often) doing at least a token system test with some kind of UNIX (usually
XENIX). Pay the few extra bucks it costs to deal with a more careful vendor.
III. Buying the basics
In this section, we cover things to look out for that are more or less
independent of price-performance tradeoffs, part of your minimum system
for running UNIX.
A. Getting Down to Cases
Cases are just bent metal. It doesn't much matter who makes those, as long as
they're above an easy minimum quality (on some *really* cheap ones, cards fail
to line up nicely with the slots, drive bays don't align with the access
cutouts, or the motherboard is ill-supported and can ground out against the
chassis). If you're fussy about RFI (Radio-Frequency Interference), it's worth
finding out whether the plastic parts of the case have conductive coating on
the inside; that will cut down emissions significantly, but a few cheap cases
omit it.
Should you buy a desktop or tower case? Our advice is go with tower unless
you're building a no-expansions personal system and expect to be using the
floppies a lot. Many vendors charge nothing extra for a tower case and the
absolute *maximum* premium I've seen is $100. What you get for that is less
desktop clutter, more and bigger bays for expansion, and often (perhaps most
importantly) a beefed-up power-supply and fan. Putting the box and its fan
under a table is good for maybe 5db off the effective noise level, too.
Airflow is also an issue; if the peripheral bays are less cramped, you get
better cooling. This is a good argument for a full tower rather than the `baby
tower' cases some vendors offer. Be prepared to buy extension cables for your
keyboard and monitor, though; vendors almost never include enough flex.
B. Power Supplies and Fans
Power supplies can matter but quality is cheap; give preference to those with a
Underwriter's Laboratories rating. There's some controversy over optimum
wattage level; on the one hand, you want enough wattage for expansion. On the
other, big supplies are noisier, and if you draw *too little* current for the
rating the delivered voltage can become unstable. And the expected wattage
load from peripherals is droppong steadily. The big old 300-watt supplies that
were designed for running several full-height 5.25" floppies are overkill in
these days of portable-driven lightweight 3.5" drives. 200 watts is good
enough these days, and the new breed of compact 200W supplies is quieter to
boot.
(About that annoying fan noise, ask if the fan on a target system has a
variable speed motor with thermostatic control --- this will cut down on noise
tremendously. If not: I have seen a rave about, but haven't used, a
thermostatic fan controller called "The Silencer". This tiny device mounts
inside your power supply and connects to the fan's power leads. It
automatically varies the fan motor speed to hold a 79 to 82F temperature.
Write Quiet Technology, Inc. PO Box 8478, Port St. Lucie FL 34985. Warning:
installing this may void your warranty!)
C. The Heart Of The Machine
Yet another basic decision, of course, is processor speed and type. Forget the
20MHz and 25MHz 386s, they're history. Right now the hot sellers in this
market are the 386/33DX and AMD 386/40DX, which I'd say are reasonable
minimum-speed engines for UNIX with X. These processors are your floor;
how far above them you want to buy depends on your budget and job mix. We'll
have much more to say about this in the section on performance tuning.
Don't bother with SX machines. Under UNIX the 16-bit bus-to-CPU path can choke
your throughput. Anyway there's not enough of a 386DX premium to matter in the
desktop market any more (laptops are a whole 'nother story). The 486SX is even
worse, a stupid marketing crock with no technical justification whatsoever.
It's a 486DX with the floating-point unit lobotomized out; the difference
*will* bite you in unobvious ways, for example if you use X which does a lot of
FP for graphics. (One respondent opined that "SX" is Intel-internal code for
"sucks".)
D. Motherboards and BIOSes
Provided you exercise a little prudence and stay out of the price basement,
motherboards and BIOS chips don't vary much in quality either. There are only
six or so major brands of motherboard inside all those cases and they're pretty
much interchangable; brand premiums are low to nonexistent and cost is strictly
tied to maximum speed and bus type. Unless you're buying from a "name" outfit
like Compaq, Dell, or AST that rolls its own motherboards and BIOSes, there are
only three major brands of BIOS chip (AMI, Phoenix, Mylex) and not much to
choose between 'em but the look of the self-test screens. One advantage UNIX
buyers have is that UNIXes are built not to rely on the BIOS code (because it
can't be used in protected mode without more pain than than it's worth). If
your BIOS will boot properly, you're usually going to be OK.
If the above sounds too rosy, there is a catch; it describes *current*
hardware, not some of the historical botches. And it's hard to know how old
what you're buying is. You might actually be buying a motherboard that's been
sitting on the dealer's back shelf for a year, with a BIOS chip in it that was
in the drawer for another year before he ever stuck it in the board. And some
of those older BIOSes and board designs are to be desperately avoided. There
have been quite a few bogus cache designs that either don't work at all
(instant panic under UNIX) or that severely degrade performance. A lot of
earlier designs have bus timing problems that show up in bad interactions with
host adapters and fancy graphics boards. Bad memory designs were also not
uncommon.
A good, tricky way to keep the vendor from shipping you these fossils is to
specify a motherboard that can take 4MB SIMs (as opposed to just the older 1MB
kind). You want to do this anyhow for functional reasons.
There are a few other potential gotchas to beware of, especially in the cheaper
off-brand boards. One is "shadow RAM", a trick some boards use for speeding up
DOS by copying the ROM contents into RAM at startup. It should be possible to
disable this. Also, on a caching motherboard, you need to be able to disable
caching in the memory areas used by expansion cards. Some cheap motherboards
fail to pass bus-mastering tests and so are useless for use with a good SCSI
interface; on others, the bus gets flaky when its turbo (high-speed) mode is
on. Fortunately, these problems aren't common.
Finally, one name-brand tip: *don't* buy DTK-brand motherboards for a UNIX
system! They generate lots of spurious interrupts, which DOS is too stupid to
be bothered by but which completely tank UNIX.
You can avoid both dangerously fossilized hardware and these little gotchas by
sticking with a system or motherboard design that's been tested with UNIX (some
help with that below).
Some other good features to look for in a motherboard include:
* Gold-plated contacts in the expansion slots and RAM sockets. Base-metal
contacts tend to grow an oxidation layer which can cause intermittent
connection faults that look like bad RAM chips or boards. (This is why, if
your hardware starts flaking out, one of the first things to do is jiggle
or remove the boards and reseat them, and press down on the RAM chips
to reseat them as well -- this may break up the oxidation layer. If
this doesn't work, rubbing what contacts you can reach with a soft
eraser is a good fast way to remove the oxidation film. Beware, some
hard erasers, including many pencil erasers, can strip off the plating, too!)
* Ability to go to 64MB on the motherboard (that is, without plug-in
daughterboards). Most EISA boards seem to have this (the popular Mylex
MAE486 board is an exception).
* The board should be speed-rated as high as your processor, of course.
It's good if it's rated higher, so upgrade to a faster processor is
just a matter of dropping in the chip and a new crystal.
If your motherboard offers multiple cache sizes, make sure you know whether the
larger cache is required when using more than a certain amount of memory. Or,
in general, fill the cache all the way -- cache-speed RAM is getting pretty
cheap.
Finally, beware the infamous FP exception bug! Some motherboards fail to
handle floating point exceptions correctly; instead of generating a SIGFPE they
lock up. The following fragment of C code will reproduce the problem:
double d;
d = 0.0;
d = 1.0 / d; /* floating divide by zero should yield SIGFPE */
John R. Levine <jo...@iecc.cambridge.ma.us> explains: "The difficulty stems
from the fact that there are two ways to handle floating exceptions on a 486,
the right way and the PC way. What the 486 wants to do is to generate an
interupt 16 when there is a floating point error, all entirely internal to the
CPU. This has been the native way to handle floating point interrupts since
the 286/287. The 286/287 and 386/387 each have a dedicated ERROR pin that the
FPU uses to tell the CPU that it's time for an error interrupt.
Unfortunately, the 8086/8087 handled interrupts differently. The error pin on
the 8087 was wired to the 8259A interrupt controller, the same interrupt
controller that handled keyboard, disk, clock, etc. interrupts. The PC/AT
enshrined IRQ 13 as the one for floating interrupts. (The details of this are
a little hazy to me, since the XT didn't have IRQ 13 tied to an 8259A, so the
AT must have at least changed the interrupt number.) PC designs have generally
wired the 287 or 387 ERROR pin to the 8259A, not to the ERROR pin on the CPU,
or at best had some poorly documented way to switch between the two interrupt
methods.
In the interest of backward compatibility, the 486 has a mode bit that says not
to handle FP exceptions automatically, but rather to freeze the FPU and send a
signal on the FERR pin, which is usually tied to an 8259A which then feeds the
interrupt back as IRQ 13. There is some extra complication involved here
because the FPU has to stay frozen until the interrupt is accepted so the CPU
can go back and look at the FPU's state. Early 386/25 chips had a bug that
would sometimes freeze up on a floating point interrupt and you had to get a
kludge socket with a PAL that fixed the timing glitch that provoked the bug.
So as likely as not, the motherboard hardware that runs FERR out and back isn't
working correctly. It's not surprising, few DOS users take floating point
seriously enough to notice whether the interrupts are working right."
When you specify a system, make clear to your vendor that the motherboard must
handle float exceptions properly. Test your motherboard's handling of
divide-by-zero; if it doesn't work, press your vendor to replace it *and
send me email*! Only by publishing a list of boards known bad can we
protect ourselves and pressure vendors to fix this problem.
The 386 UNIX Buyer's Guide posting (pc-unix/software) includes tables of
motherboards and systems known to run with various UNIX ports.
E. Peripherals
Peripherals are another matter, especially hard disks. A good rule of thumb
for balanced configurations is that the hard disk should comprise about half
(or maybe a bit more) of the total system hardware price (exception: if you're
buying a really good monitor, like 16" or over, it's going to be expensive
enough to bust this rule). Unless you're the exception who has to invoke
warranty due to a system arriving dead, most of what you buy from a dealer or
mail-order house is their ability to surf the Winchester market, make volume
buys, and burn in your disks before shipping. We'll look at disk choices in
more detail later on.
You should have a tape drive for backup, and because most UNIX vendors like to
distribute their OS on tape. Ideally, your tape backup should be able to image
your entire disk. Unfortunately, this can get *very* expensive for large
disks; 60/120MB QIC drives are fairly cheap now but larger sizes (typically
150, 250, 525 QIC tapes and 1.3gig DAT) are not (however, if you have large
drives the cost difference can quickly get weaten up by media costs). One
interesting point is that if you've gone SCSI, a 150MB QIC (comparable to the
drives now popular on Suns) may well be cheaper than older 60MB technology due
to differences in the adaptor boards required.
These days, most vendors bundle a 14" monitor and super-VGA card with 1024x768
resolution in with their systems. Details to watch are whether the card comes
loaded with 512K or 1MB of RAM (which will affect how much of that maximum
resolution and how many colors you actually get), whether the memory is
dual-ported VRAM (slightly more expensive but much faster), and whether the
monitor is interlaced or non-interlaced. The latter is better and should no
longer cost extra; look for the abbreviation NI in the ad or quote and be
suspicious if you don't see it.
We'll have much more to say about price/performance tradeoffs in peripherals
in the next major section, on performance tuning.
F. Power Protection
Finally, I strongly recommend that you buy a power conditioner to protect your
hardware. MOV-filtered power bars make nice fuses (they're cheap to replace),
but they're not enough. I've been delighted with my TrippLite 1200, which you
can get for $139 or so by mail order. A fringe benefit of this little beauty
is that if you accidently pull your plug out of the wall you may find you
actually have time to re-connect it before the machine notices!
The tecbical info in the remainder of this section is edited from material
supplied by David E. Wexelblat <d...@mtgzfs3.att.com>.
There are several levels of power protection available to the home computer
user. I break this down into 4 levels; others may have different ways of
classifying things. The levels are:
1. Surge Suppressor
2. Line Conditioners
3. Standby Power Supplies
4. Uninterruptible Power Supplies
and here's what they mean:
1. Surge suppressors
These are basically a fancy fuse between the source and your hardware; they
clamp down spikes, but can't fill in a low voltage level or dropout.
This is a bare minimum level of protection that any piece of expensive
electronics should have. Note that this applies to more than just AC power;
surge suppressors are available for (and should be used on) phone lines, and
RS-232 and parallel connections (for use on long lines; generally not needed if
the devices is colocated with the computer and all devices are protected from
outside sources). Note also that *all* devices connected to your computer need
to be protected; if you put a surge suppressor on your computer but not your
printer, then a zap on the printer may take out the computer, too.
An important fact about surge suppressors is that *they need to be replaced if
they absorb a large surge*. Besides fuses, most suppressors rely on on
components called Metal-Oxide Varistors (or MOVs) for spike suppression, which
degrade when they take a voltage hit. The problem with cheap suppressors is
that they don't tell you when the MOV is cooked, so you can end up with no
spike protection and a false sense of security --- better ones have an
indicator.
You can buy surge suppressors at any Radio Shack; for better prices, go
mail-order through Computer Shopper or some similar magazine. All of
these are low-cost devices ($10-50).
2. Line Conditioners
These devices filter noise out of AC lines. Noise can degrade your power
supply and cause it to fail prematurely. They also protect against short
voltage dropouts and include surge suppression.
My Tripp-Lite 1200 is typical of the better class of line conditioners --- a
box with a good big soft-iron transformer and a couple of moby capacitors in it
and *no* conductive path between the in and out sides. With one of these, you
can laugh at brownouts and electrical storms.
Price vary widely, from $40-400, depending on the power rating and capabilities
of the device. Mail-order from a reputable supply house is your best bet.
Line conditioners typically *don't* need to be replaced after a surge; check
to see if yours includes MOVs.
3. Standby power supplies (SPSs)
These devices are battery-based emergency power supplies that provide power for
your system via an inverter if the power fails. An SPS will generally have all
the capabilities of a line conditioner as well.
Note: these devices do not come on line until after the power fails, and have a
certain amount of delay of some of milliseconds before they come on line. If
the capacitors in your power supply are not large enough, the SPS may not cut
in in time to prevent your computer from seeing the power failure.
Note also that many SPSs are marketed as Uninterruptable Power Supplies (see
below). This is incorrect. Any device with a non-zero cutover time cannot be
a true UPS. If the ad mentions a cutover time, it's an SPS, and not a UPS.
The price range for these devices (depending largely on size and cutover time)
is $200-2000. An SPS will *not* need to be replaced after absorbing a large
surge.
4. Uninterruptable power supplies (UPSs)
These devices provide full-time isolation from the incoming AC line through a
transformer of some sort. These devices are on-line at all times, and if the
AC line fails, the batteries will cut in. Your devices will see no
interruption of their incoming AC. UPSs cost more, and provide more features.
They are the ultimate in power protection. Many UPSs have an intelligent
interface that will notify a connected device of a power failure, allowing it
to shut down cleanly. UPSs also provide the capabilities of a line
conditioner. The price range (for devices in the size range for a home
computer) are $400-$2500. An UPS will *not* need to be replaced after
absorbing a large surge.
Now, given this information, how does one decide what to get? For a system
that runs unattended, like most Unix systems, it is best to have a device that
provides both power holdover and a power failure signal. Hence, for a Unix
system, a UPS is the best idea (an SPS is not the best power protection and
most have no intelligent interface). At least one vendor sells
ordered-shutdown software for Unix, and it's fairly simple to write your own
daemon to monitor a serial port, and send init a SIGPWR signal when it sees a
certain signal.
Our recommendation for a home Unix environment is a configuration like the
following:
a) A true on-line UPS for the computer system. An intelligent
interface is mandatory, along with appropriate software for
ordered shutdown.
b) Surge suppression on all phone lines, and also on serial/parallel
lines that leave the room.
c) Line conditioners on any devices not connected to the UPS. If
you do take a power hit, it's cheaper to replace a $50 line
conditioner than a $1500 laser printer.
An important question is "How do I know how big a UPS to get?" The simple
answer is to add up the VA ratings of all devices you plan to put on it, add
~20%, then add room for growth. For a server-sized system, with a large
monitor, you probably are looking at 800-1100 VA.
One other consideration is that you typically can't put a laser printer on a
UPS. The power draw of the heater will blow the mind of the UPS. David
reports that his UPS (which normally has ~60 minutes of backup) ran out of
juice in 45 seconds after plugging in the laser printer. The other thing is
that you can't even put the laser printer on the same circuit with a UPS ---
the heater kicks on every 20-30 seconds, and most UPSs will see the current
draw as a power failure. So buy a separate line conditioner for the laser
printer.
David personally recommends surge suppressors and line conditioners from
Tripp-Lite (available both mail-order and retail), and UPSs from Best Power
Technologies (Necedah, WI - 1-(800)-356-5737). I can enthusiastically second
the TrippLite recommendation, but haven't dealt with Best Power at all. There
are many other vendors for all of these devices.
Tripp-Lite has a whole range of products, from a $10 phone-line
surge-suppressor, to line conditioners and SPSs with prces in the hundreds of
dollars. They have a line of $50-80 line conditioners that are good for most
peripherals (including your home stereo :->).
Best Power Technologies sells two lines of UPSs in the range for home systems.
The older and more expensive FERRUPS line (which is what David has) has a smart
interface, and very good filtering and surge-suppression capabilities. He says
"I have a 1.15kVA FERRUPS for my home system, which is overkill with my current
hardware (although it rode out a 45 minute power failure with nary a whisper -
no reboot). In 1990, I paid ~$1600 for this device, and that has since gone
up. They also sell a newer line of Fortress UPSs. These are better suited in
price for home systems. I don't know much about them, as they were not
available when I bought my UPS. I expect that this is what most people will
want to consider, though. In addition, Best sells Check-UPS, a software
package (in source form) for monitoring the UPS and shutting it down. I have
found Best to be a good company to deal with, with competent, knowledgable
sales people (who will be able to help you pick the right device), and helpful,
courteous, and responsive technical support."
IV. Performance tuning
Here are the places where you can trade off spending against the performance
level you want to buy and your expected job mix.
A. How to Pick your Processor
Right now, the fastest Intel 386 is 33MHz; AMD makes a 386/40. As we've
discussed, these are minimum for decent UNIX and X performance. To crank
faster than that, you have to go to a 486.
And indeed, the move to 486 processors, especially 486/33DXs, is happening very
fast. On current trends, by the end of '92 it's going to be hard to find an
Intel 386 in any new desktop machine, though the spiffier AMD clone may hang in
there longer (note that the AMD 33DX is as fast as an Intel 40DX, and the AMD
386/40DX seems to scale to sbout the equivalent of a Intel 386/50, if such a
thing actually existed).
Lots of relatively inexpensive 486/33DX systems are out there now; estimates
for the speed advantage from the 486 range from 20 to 50% with 30% pretty
widely heard (thus one of these is equivalent to about a 386/40DX). Most of
the advantage comes from the pipelining and on-chip cache. The higher level of
integration also implies better reliability. And of course, the on-chip FPU
really sizzles if you're into scientific computing or graphics.
One of Intel's most recent wrinkles is the "clock-doubler" chips. The 50DX2
runs at 25MHz externally but computes at 50MHz. A clock-doubled 33DX (compute
speed 66MHz) is sampling now, and there are persistent rumors of a
clock-doubled 50 in the works that would compute at a blistering 100MHz! Intel
likes to claim a 70% speedup for the doublers over their undoubled brethren.
I've expressed skepticism about this in previous issues, but SPECmarks suggest
that just this once the marketroids may not be lying -- much. Under UNIX, a
50DX2 is in fact nearly as fast as a true 50DX. Still, beware of anyone whose
literature passes off the DX2 qualification in the fine print; they may be
scamming about other things, too.
Right now you'll pay as much as a $1500 premium for a 486/50, as that's
relatively new technology and demands extra-fast memory to run full-out. Also,
these processors run really hot (one correspondent described the 50 as a
"toaster on a chip"). If you go this route, be sure your configuration has an
extra-heavy-duty cooling fan. Or two. And, for preference, a hefty heat sink.
Of course, if you do this you'll be ready to drop in Intel's 50MHz-external/
100MHz-internal part when it comes out later this year, and blow the doors off
all those fancy proprietary-technology workstations.
B. Of Memory In...
Buy lots of RAM, it's the cheapest way to improve real performance on any
virtual-memory system. At $30-$50 maximum per megabyte it's just plain silly
to stick with the 2-4mb now standard on most clone configurations. Go to 8,
you won't regret it; 16 if you're going to use X.
Above 16 is iffy on ISA boxes because the stock USL 4.0.3 kernel may try to do
DMA from a location the bus can't deal with. Most UNIX vendors have fixed this
by adding code that forces DMAs to take place from low memory; make absolutely
sure that includes yours before you load up beyond 16MB. The pc-unix/software
FAQ posting includes information on which vendors are known to have fixed this
problem.
C. Bus wars
Should you buy 16-bit ISA vs. 32-bit EISA? You'll pay a $600-$900 premium for
the latter. What you get in return is the ability to use things like fast
32-bit SCSI controllers and a smoother upward-migration path. On the other
hand, EISA cards are significantly more expensive. And so far, there isn't
much support for EISA-specific hardware --- a couple of vendors will drive EISA
SCSI disk and tape controllers and that's about it (of course those *are* the
most important bandwidth-eaters). All ISA cards will still work.
Of course, most of what you get from EISA is a performance boost. There are
two different theories about why EISA is better; both have their adherents.
Theory A: Bandwidth matters
UNIX has always been an I/O-intensive operating system. According to this
theory, increasing processor speed on clones can leave it spending all its time
waiting on the limited I/O capacity of the poor old 5.3MB/sec ISA bus. The
vendors all seem to think this starts at around 33MHz and that if you're buying
50MHz it definitely pays to go EISA.
Theory B: Cache is what matters
According to this theory, UNIX never comes even close to saturating the ISA-bus
bandwidth. EISA boards are faster because the premium vendors can charge for
them allows the motherboard designer more freedom and a richer parts budget.
The most important performance effect of this is that EISA boards have larger
and better-designed caches, increasing the effective memory-access speed.
There's probably some truth to both analyses. If your machine is going to
spend most of its processor time running X displays and doing other classically
compute-bound tasks, cache size matters most. On the other hand, benchmarks
show that the combination of TCP/IP and multi-user disk activity *can* saturate
ISA, and one can sometimes *see* a fast-processor machine slow down during disk
accesses...
If you're contemplating any kind of heavy-duty networking, EISA network
adapters will become rather important. A correspondent tells me he's seen
benchmarks showing what percentage of bus bandwidth is consumed by various
cards when flooding an ethernet (i.e. consuming the entire 10Mbit bandwidth of
a quiet net, as you might be when doing an FTP transfer, for instance). 8-bit
ISA cards consume 40-60% of bus bandwidth; 16-bit cards, 20-40%. 32-bit EISA
cards consume only about 5-10%. This would be particularly important in a
machine being used as a bridge, where you might be handling a large portion of
the traffic on two or more separate nets. The advantage of EISA cards may be
due to their shorter-cycle bus mastering DMA. At time of writing, only
SCO supports these, but other UNIX vendors are known to have their own drivers
in the pipeline.
D. IDE vs. SCSI
Another basic decision is IDE vs. SCSI. Either kind of disk costs about the
same, but the premium for a SCSI card varies all over the lot, partly because
of price differences between ISA and EISA SCSI cards and especially because
many motherboard vendors bundle an IDE chip right on the system board. SCSI
gives you better speed and throughput and loads the processor less, a win for
larger disks and an especially significant consideration in a multi-user
environment; also it's more expandable.
Another important win for SCSI is that it handles multiple devices much more
efficiently. If you have two IDE (or ST506 or ESDI) drives, only one can
transfer between memory and disk at once. In fact, you have to program them at
such a low level, one drive might actually be blocked from *seeking* while
you're talking to the other drive. SCSI drives are mostly autonomous and can
do everything at once; and current SCSI drives are not quite fast enough to
flood more than 1/2 the SCSI bus bandwidth, so you can have at least two drives
on a single bus pumping full speed without using it up. In reality, you don't
keep drives running full speed all the time, so you should be able to have 3-4
drives on a bus before you really start feeling bandwidth crunch.
All this having been said, don't write off IDE too quickly. Sure, it's
compatible with the nasty old ST506 interface, but it's *much* faster. It
remains the cost-effective choice for smaller drives (up to 500MB) on systems
that won't be hitting the disk constantly. Unless you're running a heavily
used network or database server, don't assume SCSI will make any noticeable
difference.
(If you can do your own installation, I hear that used 150/250MB SCSI drives
are getting quite common and cheap on the net. All 150MB QIC type drives can
do 250MB on extended-length tapes, though some manufacturers discourage you
from doing this to avoid excessive heade wear. But back to disks...)
The following, by Ashok Singhal <ash...@duckjibe.eng.sun.com> of Sun
Microsystems, is a valiant attempt to demystify SCSI terminology.
The terms "SCSI" and "SCSI-2" refer to two different specifications.
Each specification has a number of options. Many of these options are
*independent* of each other. I like to think of the main options (there are
others that I'll skip over because I don't know enough about them to talk
about them on the net) by classifying them into five categories:
1. Logical
This refers to the commands that the controllers understand.
SCSI-2 defined a common cammand set that is pretty much a
superset of the SCSI command set.
2. Data Width
8 bits (+ 1 parity) -> "normal"
16-bits (+ 2 parity) -> "wide"
32-bits (+ 4 parity) -> I don't know, "extra-wide??"
All three options are available in SCSI-2 (yes,
the draft spec I have even shows 32-bits!), although
8-bit wide is still by far the most common. Not sure, but I believe
SCSI defined only 8-bit wide data path.
3. Electrical Interface
single-ended (max cable length 6 meters)
differential (max cable length 25 meters)
Both options are available for SCSI-2 (I'm not sure about SCSI,
but I think both options were available also)
and this option is independent of options 2, 4, 5. Differential
is less common but allows better noise immunity and longer
cables.
4. Handshake
Asynchronous (requests and acks alternate)
Synchronous (multiple requests can be outstanding)
Both options are available for SCSI-2 (Not sure about SCSI,
but I think both were available also). This is negotiated
between each target and initiator; asynchronous and synchronous
transfers can occur on the same bus. This is independent of
2, 3 (Not sure about 1).
5. Synchronous Speed (does not apply for asynchronous option)
"Normal" is up to 5 Mtransfers/sec ( = 5MB/s for 8-bit wide, more
for wider)
"Fast" is up to 10 Mtransfers/s ( = 10 MB/s for 8-bit wide, more
for wider)
The fast option is defined only in SCSI-2.
This options basically defines shorter timing parameters
such as the assertion period and hold time.
The parameters of the synchronous transfer are negotiated
between each target and initiator so different speed transfers
can occur over the same bus.
E. Other Disk Decisions
Look at seek times and transfer rates for your disk; under UNIX disk speed and
throughput are so important that a 1-millisecond difference in average seek
time can be noticeable.
Previous issues said "Disk caching is good, but there can be too much of a
good thing. Excessively large caches will slow the system because the overhead
for cache fills swamps the real accesses (this is especially a trap for
databases and other applications that do non-sequential I/O). More than 100K
of cache is probably a bad idea for a general-purpose UNIX box; watch out for
manufacturers who inflate cache size because memory is cheap and they think
customers will be impressed by big numbers." This may no longer be true on
current hardware; in particular, most controllers will interrupt a cache-fill
to fulfill a `real' read request.
However, it may be that *any* hardware disk caching is a lose for UNIX! Scott
Bennett <benn...@mp.cs.niu.edu> reports a discussion on comp.unix.wizards:
"nobody found the hardware disk caches to be as effective in terms of
performance as the file system buffer cache...In many cases, disabling the
hardware cache improved system performance substantially. The interpretation
of these results was that the caching algorithm in the kernel was superior to,
or at least better tuned to UNIX accesses than, the hardware caching
algorithms."
Thus, if your disk controller allows it, try disabling the cache. Your
throughput may go up!
F. Souping Up X Performance
One good way to boost your X performance is to invest in a graphics card with a
dedicated blitter or high-speed local-bus connection, like the ATI series or
the S3-based Quantum, Wind/X and Orchid Fahrenheit 1280. A number of clone
vendors offer these accelerator options relatively cheap (about $300) and can
make your X go like a banshee; however, stock X doesn't support them yet ---
and the third-party servers that do (such as MetroLink's or SGCS's) ain't
cheap (SGCS's is about $300, MetroLink's about $900).
S3 is an interesting phenomenon. Though several vendors advertise their S3
connector as "proprietary" it's actually an anticipation of a forthcoming VESA
(Video Electronics Standards Association) standard. S3 cards are thus
semi-interchangeable.
In general, the ATI approach (normal bus, dedicated blitter and optimization
for special functions like character drawing) will speed up text display, text
scrolling and window resize/move operations a lot, but line-drawing and
graphics only a little. S3, on the other hand, speeds up high-bandwidth
graphics drawing a lot but doesn't have as big an advantage for ordinary
text operations. You pays your money and takes your choice. Benchmarks
indicate that most non-CAD users are better served by the ATI approach.
If you're feeling *really* flush, plump for a 15", 17" or even 20" monitor.
The larger size can make a major difference in viewing comfort. Also you'll be
set for VESA 1280x1024 when everybody gets to supporting that. In the mean
time, the bigger screen will allow you to use fonts in smaller pixel sizes so
that your text windows can be larger, giving you a substantial part of the
benefit you'd get from higher pixel resolutions.
V. Of Mice and Machines
In a previous issue, I claimed that all mice and trackballs are the same for
compatibility purposes. I was wrong -- seriously wrong. The more I found out,
the messier the picture gets. The following is an attempt to sort out all the
confusion. Thanks to Jim McCarthy at Logitech for digging into the matter
and somewhat alleviating my ignorance.
Mice and trackballs used to be simple; now, thanks to Microsoft, they're
complicated. In the beginning, there was only the Mouse Systems 3-button
serial mouse; this reported status to a serial port 30 times a second using a
5-byte serial packet encoding now called "C" protocol. The Logitech Series 7
and 9 mice were Mouse Systems-compatible. All UNIXes that have any mouse
support at all understand C-protocol serial mice.
Then Microsoft got into the act. They designed a two-button serial mouse which
reports only deltas in a three-byte packet; that is, it sends changes in button
status and motion reports only when the mouse is actually moving. This is
called `M' protocol. Microsoft sold a lot of mice, so Logitech switched from
`C' to `M' --- but they added a third button, state changes for which show up
in an optional fourth byte. Thus, `M+' protocol, upward-compatible with
Microsoft's `M'. Most UNIX vendors add support for M+ mice, but it's wise to
check.
Bus mice are divided into 8255 and InPort types. These report info
continuously at 30 or 60 Hz (though InPort mice have an option for reporting
deltas only), and you get interrupts on events and then have to poll hardware
ports for details. More on these next issue.
In addition to serial mice and bus mice, there are "keyboard mice". On PS/2s
there are two identical-looking keyboard ports, labeled (with icons) "mouse" &
"keyboard". Both are 8 or 9 pin mini-DIN's that look like the regular PC
keyboard port only smaller. I don't know what logical protocol the keyboard
mouse speaks. Physically, the connector is eventually connected to the
keyboard processor (often an 8042). The same keyboard processor that decodes
the keyboard decodes the mouse. PS/2s have this port, many newer ISA/EISA
motherboards do as well.
All things considered, UNIX users are probably best off going with a serial
mouse (most current clone motherbords give you two serial ports, so you can
dedicate one to this and still have one for the all-important modem). Not only
are the compatibility issues less daunting, but a serial mouse loads the
multitasking system less due to interrupt frequency. Beware that most clone
vendors, being DOS oriented, bundle M-type mice for which UNIX support is
presently spotty, and they may not work with your X. Ignore the adspeak about
dpi and pick a mouse/trackball that feels good to your hand.
VI. When, Where and How to Buy
If you're a serious UNIX hacker for either fun or profit, you're probably in
the market for what the mail-order vendors think of as a high-end or even
`server' configuration, and you're going to pay a bit more than the DOS
lemmings. On the other hand, prices keep dropping, so there's a temptation to
wait indefinately to buy. A tactic that makes a lot of sense in this market,
if you have the leisure, is to fix in your mind a configuration and a trigger
price that's just a little sweeter than the market now offers and buy when
that's reached.
Direct-mail buying makes a lot of sense today for anyone with more technical
savvy than J. Random Luser in a suit. Even from no-name mail-order houses,
parts and system quality tend to be high and consistent, so conventional
dealerships don't really have much more to offer than a warm fuzzy feeling.
Furthermore, competition has become so intense that even mail-order vendors
today have to offer not just lower prices than ever before but warranty and
support policies of a depth that would have seemed incredible a few years back.
For example, many bundle a year of on-site hardware support with their medium-
and high-end "business" configurations for a very low premium over the bare
hardware.
Note, however, that assembling a system yourself out of mail-order parts is
*not* likely to save you money over dealing with the mail-order systems
houses. You can't buy parts at the volume they do; the discounts they command
are bigger than the premiums reflected in their prices. The lack of any
system warranty or support can also be a problem even if you're expert enough
to do the integration yourself --- because you also assume all the risk of
defective parts and integration problems.
Cruise through "Computer Shopper" and similar monthly ad compendia. Even if
you decide to go with a conventional dealer, this will tell you what *their*
premiums look like.
Another alternative to conventional dealerships (with their designer "looks",
stone-ignorant sales staff, and high overheads that *you* pay for) is to go
with one of the thousands of the hole-in-the-wall stores run by immigrants from
the other side of the International Date Line. They're usually less ignorant
and have much lower overheads; they do for you locally what a mail-order house
would, that is assemble and test parts they get for you from another tier of
suppliers. You won't get plush carpeting or a firm handshake from a white guy
with too many teeth and an expensive watch, but then you didn't really want to
pay for those anyway, right?
A lot of vendors bundle DOS 5.0 and variable amounts of DOS apps with their
hardware. You can tell them to lose all this cruft and they'll shave $50 or
$100 off the system price. However, David Wexelblat observes "there are at
least two situations in which the Unix user will need DOS available: 1) most,
if not all, EISA configuration utilities run under DOS, and 2) SCSICNTL.EXE by
Roy Neese is a godsend for dealing with SCSI devices on Adaptec boards."
Don't forget that (most places) you can avoid sales tax by buying from an
out-of-state mail-order outfit, and save yourself 6-8% depending on where you
live. If you live near a state line, buying from a local outfit you can often
win, quite legally, by having the stuff shipped to a friend or relative just
over it. Best of all is a buddy with a state-registered dealer number; these
aren't very hard to get and confer not just exemption from sales tax but
(often) whopping discounts from the vendors. Hand him a dollar afterwards to
make it legal.
(Note: I have been advised that you shouldn't try the latter tactic in
Florida -- they are notoriously tough on "resale license" holders).
(Note II: The Supreme Court recently ruled that states may not tax out-of-state
businesses under existing law, but left the way open for Congress to pass
enabling legislation. Let's hope the mail-order industry has good lobbyists.)
On the other hand, one good argument for buying locally is that you may have to
pay return postage if you ship the system back. On a big, heavy system, this
can make up the difference from the savings on sales tax.
VII. Questions You Should Always Ask Your Vendor
A. Minimum Warranty Provisions
The weakest guarantee you should settle for in the mail-order market should
include:
* 72-hour burn-in to avoid that sudden infant death syndrome. (Also,
try to find out if they do a power-cycling test and how many repeats
they do; this stresses the hardware much more than steady burn-in.)
* 30 day money-back guarantee. Watch out for fine print that weakens this
with a restocking fee or limits it with exclusions.
* 1 year parts and labor guarantee (some vendors give 2 years).
* 1 year of 800 number tech support (many vendors give lifetime support).
Additionally, many vendors offer a year of on-site service free. You should
find out who they contract the service to. Also be sure the free service
coverage area includes your site; some unscrupulous vendors weasel their way
out with "some locations pay extra", which translates roughly to "through the
nose if you're further away than our parking lot".
If you're buying store-front, find out what they'll guarantee beyond the
above. If the answer is "nothing", go somewhere else.
B. Documentation
Ask your potential suppliers what kind and volume of documentation they supply
with your hardware. You should get, at minimum, operations manuals for the
motherboard and each card or peripheral; also an IRQ list, and a bad-block
listing if your Winchester is ESDI rather than IDE or SCSI (the latter two
types of drive do their own bad-block mapping internally). Skimpiness in this
area is a valuable clue that they may be using no-name parts from Upper
Baluchistan, which is not necessarily a red flag in itself but should prompt
you to ask more questions.
C. A System Quality Checklist
There are various cost-cutting tactics a vendor can use which bring down the
system's overall quality. Here are some good questions to ask:
* Is the memory zero-wait-state? One or more wait states allows the vendor to
use slower and cheaper memory but will slow down your actual memory subsystem
throughput. This is a particularly important question for the *cache*
memory!
* Is the monitor non-interlaced? Does it have a tilt-and-swivel base? Is it
*color*? Yes, if you don't see it in the ad, ask; some lowball outfits will
try to palm off so-called "black & white VGA" monitors on you. What's the
vertical scan rate? 60Hz is SVGA standard; 72Hz is VESA standard and
minimal for flicker-free operation; 80Hz is cutting-edge. What's the dot
pitch? .31mm is minimal, .28mm or .27mm is good. You need .28mm for X.
A slightly larger dot pitch is acceptable in a larger monitor (16" or more).
* If you're buying a factory-configured system, does it have FCC certification?
While it's not necessarily the case that a non-certified system is going
to spew a lot of radio-frequency interference, certification is legally
required --- and becoming more important as clock frequencies climb. Lack
of that sticker may indicate a fly-by-night vendor, or at least one in
danger of veing raided and shut down!
VIII. Things to Check when Buying Mail-Order
A. Tricks and Traps in Mail-Order Warranties
Reading mail-order warranties is an art in itself. A few tips:
Beware the deadly modifier "manufacturer's" on a warranty; this means you have
to go back to the equipment's original manufacturer in case of problems and
can't get satisfaction from the mail-order house. Also, manufacturer's
warranties run from the date *they* ship; by the time the mail-order house
assembles and ships your system, it may have run out!
Watch for the equally deadly "We do not guarantee compatibility". This gotcha
on a component vendor's ad means you may not be able to return, say, a video
card that fails to work with your motherboard.
Another dangerous phrase is "We reserve the right to substitute equivalent
items". This means that instead of getting the high-quality name-brand parts
advertised in the configuration you just ordered, you may get those no-name
parts from Upper Baluchistan --- theoretically equivalent according to the
spec sheets, but perhaps more likely to die the day after the warranty expires.
Substitution can be interpreted as "bait and switch", so most vendors are
scared of getting called on this. Very few will hold their position if you
press the matter.
Another red flag: "Only warranted in supported environments". This may mean
they won't honor a warranty on a non-DOS system at all, or it may mean they'll
insist on installing the UNIX on disk themselves.
One absolute show-stopper is the phrase "All sales are final". This means you
have *no* options if a part doesn't work. Avoid any company with this policy.
B. Special Questions to Ask Mail-Order Vendors Before Buying
* Does the vendor pay for shipping? What's the delivery wait?
* If you need to return your system, is there a restocking fee? and will the
vendor cover the return freight? Knowing the restocking fee can be
particularly important, as they make keep you from getting real satisfaction
on a bad major part. Avoid dealing with anyone who quotes more than a 15%
restocking fee --- and it's a good idea, if possible, to avoid any dealer
who charges a restocking fee at all.
C. Payment Method
It's a good idea to pay with AmEx or Visa or MasterCard; that way you can
stop payment if you get a lemon, and may benefit from a buyer-protection plan
using the credit card company's clout. However, watch for phrases like "Credit
card surcharges apply" or "All prices reflect 3% cash discount" which mean
you're going to get socked extra if you pay by card.
VIII. Which Clone Vendors to Talk To
I went through the March 1992 issue of Computer Shopper calling vendor 800
numbers with the following question: "Does your company have any
configurations aimed at the UNIX market; do you use UNIX in-house; do
you know of any of the current 386 or 486 ports running successfully
on your hardware?
I didn't call vendors who didn't advertise an 800 number. This was only partly
to avoid phone-bill hell; I figured that toll-free order & info numbers are so
standard in this industry sector that any outfit unable or unwilling to spring
for one probably couldn't meet the rest of the ante either. I also omitted
parts houses with token systems offerings and anybody who wasn't selling
desktops or towers with a 386/33DX or heavier processor inside.
After plundering Computer Shopper, I called up a couple of "name" outfits that
don't work direct-mail and got the same info from them.
The answers I get revealed that for most clone vendors UNIX is barely a blip on
the screen. Only a few have tested with an SVr4 port. Most seem barely aware
that the market exists. Many seem to rely on their motherboard vendors to tell
them what they're compatible, without actually testing whole systems. Since
most compatibility problems have to do with peripheral cards, this is a
problem.
Here's a summary of the most positive responses I got:
A --- Advertises UNIX compatibility.
C --- Has known UNIX customers.
I --- Uses UNIX in-house.
T --- Have formally tested UNIX versions on their hardware.
F --- Have 486/50 systems
* --- Sounded to me like they might actually have a clue about the UNIX market.
Vendor A C I T F * Ports known to work
--------------- - - - - - - -----------------------------------------------
ARC . . X X . . SCO XENIX 2.3.2, SCO UNIX 3.2.1
AST . X X X X * SCO UNIX 3.2.4, ODT 2.0 Microport V/4
Allegro . . X X . . SCO XENIX 3.2.4
Altec . X . X . . XENIX (no version given).
Ares . X X X X * AT&T 3.2, ISC (version unknown)
Basic Time . X X X X * SCO XENIX 2.3.2, have in-house UNIX experts.
Binary Tech . X . X X . Claims to work with all versions.
Blue Dolphin . X . X X * SCO XENIX.
CCSI X X . . X . They've used SCO XENIX, no version given.
CIN . X . . . . SCO UNIX (version not specified)
CSS . X . X . * SCO 3.2.2, ISC 3.0, SCO ODT. See Will Harper.
Centrix X . . . . . No specifics on versions.
Compudyne . X X X X . Couldn't get details on which versions.
Comtrade . X . X X . Couldn't get details on which versions.
Datom X X X X X . SCO XENIX 3.2.
Dell X X X X X * See Dell SVr4 data.
Desert Sands X X . X X . SCO UNIX 3.2.4
Digitech . X . X . . SCO UNIX 3.2.1, XENIX 2.3.1
EPS X X X X . * SCO XENIX 3.2.4, ISC & AT&T (versions not sp.)
Gateway 2000 X X X X X * SCO UNIX 3.2.0. XENIX 2.3.4 ISC 3.0, ESIX 4.0.3
HD Computer . X . X X . SCO UNIX 3.2, SCO XENIX 3.2.2
HiQ . X . X . . SCO UNIX (version not specified)
Infiniti . X . X X . SCO UNIX (versions not specified)
Insight . . X . X . SCO XENIX 3.2.4. No tech support for UNIX
Keydata X . X X X * SCO version 4, ISC 3.2
Legatech . X . . X . SCO UNIX, ISC (versions not specified)
MicroGeneration . . X . . . Uses XENIX.
MicroLab X . . . . . SCO UNIX, SCO XENIX
MicroSmart X X . X . . SCO XENIX (version not specified)
Microlink X . . X X . SCO XENIX (version not specified)
Myoda X X . X X . SCO XENIX 3.2.2, ISC 3.2
Naga . X . X X * SCO & XENIX 3.2.
Northgate X X . X X * SCO UNIX 3.2
PC Brand . X X X . . SCO XENIX, ISC UNIX
PC Professional . X . X . . ISC 3.2
PC-USA X X . X . . ISC 5.3.2 and SCO 3.2
Profex . X . X . . SCO XENIX 3.2.
Royal Computer . X . . X . No details on versions.
SAI X X . X X . SCO UNIX 3.2.2.
Santronics . . X X X . SCO XENIX 3.2.4
Solidtech . X . . . . Dell (no version given), ISC 3.2.
Strobe . . . X X . SCO, Microport, ISC (no version numbers given)
Swan X X X X X * SCO 2.3.1, UNIX 3.2, ISC 3.2v2.0.2
TriStar . X X X X * SCO UNIX 3.2.2, XENIX 2.3.2, ISCr4
Zenon . X . X X * SCO UNIX (version not specified)
Zeos . X X X X * SCO XENIX 3.2.4, AT&T 3.2
Special notes about a few vendors who appear to have a clue:
Ares targets some of its systems for UNIX CAD use. They have a house wizard
name Ken Cooper (everybody calls him "K.C.").
EPS targets some 486 EISA configurations for UNIX.
Swan doesn't know the UNIX market very well yet, but their project manager
wants a bigger piece of it and is interested in doing some of the right
things. They have a house wizard, one John Buckwalter.
Dell, of course, supports an industry-leading SVr4 port. They're a bit on
the pricy side, but high quality and very reliable. Lots of UNIX expertise
there; some of it hangs out on the net.
Zeos is on the net as zeos.com, with a uunet connection; they host a UNIX
BBS. They have an in-house UNIX group; talk to Ken Germann for details. There
are biz.zeos.general and biz.zeos.announce groups on USENET.
Special notes about a lot of vendors who appear to have *no* clue:
Vendors where I couldn't get a real person on the line, either because
no one answered the main number or because I couldn't raise anyone at
tech support after being directed there: Sunnytech, Quantex, AMS, USA
Flex, Lapine, Syntax Computer, MicroTough, PAC International, The Portable
Warehouse.
Vendors where the question met with blank incomprehension, puzzlement,
consternation, or "We've never tested with UNIX": Allur, AmtA, Aplus, HiTech,
Locus Digital Products, LodeStar, TriStar Computers, Ultra-Comp, UTI Computers,
PC Turbo Corp, Evertek, Microcomputer Concepts, Jinco Computers, UWE, ToughCom,
System Dynamics Group, Terribly Fast Bus Systems.
Vendors who understood the questions but had no answer: Bulldog Computer
Products, LT Plus, Standard Computer, JCC.
Vendors who said "Yes, we're UNIX-compatible" but had no details of any tests:
CompuCity.
Vendors who said "Go ask our motherboard vendor": Ariel Design, Lucky Computer
Co., V-com, Professional Computer, MicroLine, MileHi.
Vendors who sent me to a toll number: Absec, Hokkins, New Technologies, Mirage.
Vendors that believe they have UNIX customers, but can't be any definite than
that: Austin Computer Systems, PC Professional, Treasure Chest Computer
Systems, CompuAdd Express, FastMicro, MidWest Micro.
Final note:
If you order from these guys, be sure to tell them you're a UNIX customer
and don't need the bundled DOS. This will shave some bucks off the system
price, *and* it may encourage them to pay more attention to the UNIX market.
--
Send your feedback to: Eric Raymond = e...@snark.thyrsus.com
Path: sparky!uunet!cbmvax!snark!esr
From: e...@snark.thyrsus.com (Eric S. Raymond)
Newsgroups: comp.unix.sysv386,comp.sys.intel,news.answers
Subject: PC-Clone UNIX Hardware Buyer's Guide
Message-ID: <1hQSvZ#5NQTrw8sLg8W9wJLZM5ny9ts=esr@snark.thyrsus.com>
Date: 12 Aug 92 17:23:18 GMT
Expires: 5 Nov 92 00:00:00 GMT
Sender: e...@snark.thyrsus.com (Eric S. Raymond)
Followup-To: comp.unix.sysv386
Distribution: world
Lines: 1283
Summary: Tips on how and where to buy hardware for your UNIX.
Approved: news-answers-requ...@MIT.Edu
Archive-name: pc-unix/hardware
Last-update: Wed Aug 12 13:15:22 1992
Version: 6.0
What's new in this issue:
* How you should be careful about SIMM sizes
* Benchmark data on 486 vs, 386
* Price lists for MetroLink & SGCS X products
* More about memory and disk caches.
[My apologies to all for being a week late. Press of work and all that...]
Gentle Reader: if you end up buying something based on information from this
Guide, please do yourself and the net a favor; make a point of telling the
vendor "Eric's FAQ sent me" or some equivalent. The idea isn't to hype me
personally, I've already got all the notoriety I need from doing things like
_The_New_Hacker's_Dictionary_ --- but if we can show vendors that the Guide
influences a lot of purchasing decisions, I can be a more powerful advocate for
the net's interests, and for you.
0. Contents
I. Introduction
II. Overview of the Market
III. Buying the basics
A. Getting Down to Cases
B. Power Supplies and Fans
C. The Heart Of The Machine
D. Motherboards and BIOSes
E. Peripherals
F. Power Protection
IV. Performance tuning
A. How To Pick Your Processor
B. Of Memory In...
C. Bus Wars
D. IDE vs. SCSI
E. Other Disk Decisions
F. Souping Up X Performance
V. Of Mice and Machines
VI. When, Where and How to Buy
VII. Questions You Should Always Ask Your Vendor
A. Minimum Warranty Provisions
B. Documentation
C. A System Quality Checklist
VIII. Things to Check when Buying Mail-Order
A. Tricks and Traps in Mail-Order Warranties
B. Special Questions to Ask Mail-Order Vendors Before Buying
C. Payment Method
IX. Which Clone Vendors to Talk To
I. Introduction
The purpose of this posting is to give you the background information you need
to be a savvy buyer of 386/486 hardware for running UNIX. It is aimed
especially at hackers and others with the technical skills and confidence to go
to the mail-order channel, but contains plenty of useful advice for people
buying store-front retail. It was formerly part of 386-buyers-faq issues 1.0
through 4.0, and is still best read in conjunction with the pc-unix/software
FAQ descended from that posting.
This document is maintained and periodically updated as a service to the net by
Eric S. Raymond <e...@snark.thyrsus.com>, who began it for the very best
self-interested reason that he was in the market and didn't believe in plonking
down several grand without doing his homework first (no, I don't get paid for
this, though I have had a bunch of free software and hardware dumped on me as a
result of it!). Corrections, updates, and all pertinent information are
welcomed at that address. The editorial `we' reflects the generous
contributions of many savvy USENETters.
This posting is periodically broadcast to the USENET group comp.unix.sysv386
and to a list of vendor addresses. If you are a vendor representative, please
check to make sure any information pertaining your company is current and
correct. If it is not, please email me a correction ASAP. If you are a
hardware-knowledgeable user, please send me any distillation of your
experience that you think might improve this posting.
II. Overview of the Market
The central fact about 386/486 clone hardware that conditions every aspect of
buying it is this: more than anywhere else in the industry, de-facto hardware
standards have created a commodity market with low entry barriers, lots of
competitive pressure, and volume high enough to amortize a *lot* of development
on the cheap.
The result is that this hardware gives you lots of bang-per-buck, and it's
getting both cheaper and better all the time. Furthermore, margins are thin
enough that vendors have to be lean, hungry, and *very* responsive to the
market to survive. You can take advantage of this, but it does mean that much
of the info in the rest of this document will be stale in three months and
completely obsolete in six.
One good general piece of advice is that you should avoid both the highest-end
new-technology systems (those not yet shipping in volume) and the very cheapest
systems put out by vendors competing primarily on price. The problem with
the high end is that it usually carries a hefty "prestige" price premium, and
may be a bit less reliable on average because the technology hasn't been
through a lot of test/improve cycles. The problem with the low end is that
price-cutters sometimes settle for marginal components. UNIX is more
sensitive to hardware flakiness than DOS, so cut-price systems that might
deliver consistently for DOS lemmings can come around and bite you. Use a
little care, and spend the $200-$300 to stay out of the basement. It's worth
it.
The last point deserves a little amplification. In the PC world, there's a lot
of "if it doesn't fail, it's OK". It is common to ignore normal engineering
tolerances --- the allowances for variations in components, temperature,
voltage margins, and the like --- and to assume that anything which doesn't
fail outright must work. Watch out! For example, The ISA bus was originally
designed for 6 MHz. IBM later updated that to 8 MHz, and that's as much of a
standard as there is, yet there are motherboards that will let you (try to!)
run it at 12 MHz --- 50% over spec. Some cards are actually designed to work
at that speed with proper tolerances. Others might work...or they might flake
out when they get warm. Any systems vendor above the fly-by-night level is
going to shoot for a little more reliability than this, burning in systems and
(often) doing at least a token system test with some kind of UNIX (usually
XENIX). Pay the few extra bucks it costs to deal with a more careful vendor.
III. Buying the basics
In this section, we cover things to look out for that are more or less
independent of price-performance tradeoffs, part of your minimum system
for running UNIX.
A. Getting Down to Cases
Cases are just bent metal. It doesn't much matter who makes those, as long as
they're above an easy minimum quality (on some *really* cheap ones, cards fail
to line up nicely with the slots, drive bays don't align with the access
cutouts, or the motherboard is ill-supported and can ground out against the
chassis). If you're fussy about RFI (Radio-Frequency Interference), it's worth
finding out whether the plastic parts of the case have conductive coating on
the inside; that will cut down emissions significantly, but a few cheap cases
omit it.
Should you buy a desktop or tower case? Our advice is go with tower unless
you're building a no-expansions personal system and expect to be using the
floppies a lot. Many vendors charge nothing extra for a tower case and the
absolute *maximum* premium I've seen is $100. What you get for that is less
desktop clutter, more and bigger bays for expansion, and often (perhaps most
importantly) a beefed-up power-supply and fan. Putting the box and its fan
under a table is good for maybe 5db off the effective noise level, too.
Airflow is also an issue; if the peripheral bays are less cramped, you get
better cooling. This is a good argument for a full tower rather than the `baby
tower' cases some vendors offer. Be prepared to buy extension cables for your
keyboard and monitor, though; vendors almost never include enough flex.
B. Power Supplies and Fans
Power supplies can matter but quality is cheap; give preference to those with a
Underwriter's Laboratories rating. There's some controversy over optimum
wattage level; on the one hand, you want enough wattage for expansion. On the
other, big supplies are noisier, and if you draw *too little* current for the
rating the delivered voltage can become unstable. And the expected wattage
load from peripherals is droppong steadily. The big old 300-watt supplies that
were designed for running several full-height 5.25" floppies are overkill in
these days of portable-driven lightweight 3.5" drives. 200 watts is good
enough these days, and the new breed of compact 200W supplies is quieter to
boot.
(About that annoying fan noise, ask if the fan on a target system has a
variable speed motor with thermostatic control --- this will cut down on noise
tremendously. If not: I have seen a rave about, but haven't used, a
thermostatic fan controller called "The Silencer". This tiny device mounts
inside your power supply and connects to the fan's power leads. It
automatically varies the fan motor speed to hold a 79 to 82F temperature.
Write Quiet Technology, Inc. PO Box 8478, Port St. Lucie FL 34985. Warning:
installing this may void your warranty!)
C. The Heart Of The Machine
Yet another basic decision, of course, is processor speed and type. Forget the
20MHz and 25MHz 386s, they're history. Right now the hot sellers in this
market are the 386/33DX and AMD 386/40DX, which I'd say are reasonable
minimum-speed engines for UNIX with X. These processors are your floor;
how far above them you want to buy depends on your budget and job mix. We'll
have much more to say about this in the section on performance tuning.
Don't bother with SX machines. Under UNIX the 16-bit bus-to-CPU path can choke
your throughput. Anyway there's not enough of a 386DX premium to matter in the
desktop market any more (laptops are a whole 'nother story). The 486SX is even
worse, a stupid marketing crock with no technical justification whatsoever.
It's a 486DX with the floating-point unit missing; the difference *will* bite
you in unobvious ways, for example if you use X which does a lot of FP for
graphics. (One respondent opined that "SX" is Intel-internal code for
"sucks".)
D. Motherboards and BIOSes
Provided you exercise a little prudence and stay out of the price basement,
motherboards and BIOS chips don't vary much in quality either. There are only
six or so major brands of motherboard inside all those cases and they're pretty
much interchangable; brand premiums are low to nonexistent and cost is strictly
tied to maximum speed and bus type. Unless you're buying from a "name" outfit
like Compaq, Dell, or AST that rolls its own motherboards and BIOSes, there are
only four major brands of BIOS chip (AMI, Phoenix, Mylex, Award) and not much
to choose between 'em but the look of the self-test screens. One advantage
UNIX buyers have is that UNIXes are built not to rely on the BIOS code (because
it can't be used in protected mode without more pain than than it's worth). If
your BIOS will boot properly, you're usually going to be OK.
If the above sounds too rosy, there is a catch; it describes *current*
hardware, not some of the historical botches. And it's hard to know how old
what you're buying is. You might actually be buying a motherboard that's been
sitting on the dealer's back shelf for a year, with a BIOS chip in it that was
in the drawer for another year before he ever stuck it in the board. And some
of those older BIOSes and board designs are to be desperately avoided. There
have been quite a few bogus cache designs that either don't work at all
(instant panic under UNIX) or that severely degrade performance. A lot of
earlier designs have bus timing problems that show up in bad interactions with
host adapters and fancy graphics boards. Bad memory designs were also not
uncommon.
A good, tricky way to keep the vendor from shipping you these fossils is to
specify a motherboard that can take 4MB SIMMs (as opposed to just the older 1MB
kind). You want to do this anyhow for functional reasons.
There are a few other potential gotchas to beware of, especially in the cheaper
off-brand boards. One is "shadow RAM", a trick some boards use for speeding up
DOS by copying the ROM contents into RAM at startup. It should be possible to
disable this. Also, on a caching motherboard, you need to be able to disable
caching in the memory areas used by expansion cards. Some cheap motherboards
fail to pass bus-mastering tests and so are useless for use with a good SCSI
interface; on others, the bus gets flaky when its turbo (high-speed) mode is
on. Fortunately, these problems aren't common.
Finally, one name-brand tip: *don't* buy DTK-brand motherboards for a UNIX
system! They generate lots of spurious interrupts, which DOS is too stupid to
be bothered by but which completely tank UNIX.
You can avoid both dangerously fossilized hardware and these little gotchas by
sticking with a system or motherboard design that's been tested with UNIX (some
help with that below).
Some other good features to look for in a motherboard include:
* Gold-plated contacts in the expansion slots and RAM sockets. Base-metal
contacts tend to grow an oxidation layer which can cause intermittent
connection faults that look like bad RAM chips or boards. (This is why, if
your hardware starts flaking out, one of the first things to do is jiggle
or remove the boards and reseat them, and press down on the RAM chips
to reseat them as well -- this may break up the oxidation layer. If
this doesn't work, rubbing what contacts you can reach with a soft
eraser is a good fast way to remove the oxidation film. Beware, some
hard erasers, including many pencil erasers, can strip off the plating, too!)
* Ability to go to 64MB on the motherboard (that is, without plug-in
daughterboards). Most EISA boards seem to have this (the popular Mylex
MAE486 board is an exception).
* The board should be speed-rated as high as your processor, of course.
It's good if it's rated higher, so upgrade to a faster processor is
just a matter of dropping in the chip and a new crystal.
If your motherboard offers multiple cache sizes, make sure you know whether the
larger cache is required when using more than a certain amount of memory. Or,
in general, fill the cache all the way -- cache-speed RAM is getting pretty
cheap.
Note, however, that hardware caches for system boards are really designed to
achieve effective 0 wait state status, rather than perform any significant
buffering of data. As a general rule applicable to all clones, 64Kb cache
handles up to 16Mb memory. 256Kb cache handles up to 64Mb. 128Kb cache is
redundant for a 16MB system, since additional caching will not give you any
benefit. This means tthat running with 8Mb RAM, there is no difference between
a 64Kb, 128Kb, or 256Kb cache on the systemboard.
Finally, beware the infamous FP exception bug! Some motherboards fail to
handle floating point exceptions correctly; instead of generating a SIGFPE they
lock up. The following fragment of C code will reproduce the problem:
double d;
d = 0.0;
d = 1.0 / d; /* floating divide by zero should yield SIGFPE */
John R. Levine <jo...@iecc.cambridge.ma.us> explains: "The difficulty stems
from the fact that there are two ways to handle floating exceptions on a 486,
the right way and the PC way. What the 486 wants to do is to generate an
interupt 16 when there is a floating point error, all entirely internal to the
CPU. This has been the native way to handle floating point interrupts since
the 286/287. The 286/287 and 386/387 each have a dedicated ERROR pin that the
FPU uses to tell the CPU that it's time for an error interrupt.
Unfortunately, the 8086/8087 handled interrupts differently. The error pin on
the 8087 was wired to the 8259A interrupt controller, the same interrupt
controller that handled keyboard, disk, clock, etc. interrupts. The PC/AT
enshrined IRQ 13 as the one for floating interrupts. (The details of this are
a little hazy to me, since the XT didn't have IRQ 13 tied to an 8259A, so the
AT must have at least changed the interrupt number.) PC designs have generally
wired the 287 or 387 ERROR pin to the 8259A, not to the ERROR pin on the CPU,
or at best had some poorly documented way to switch between the two interrupt
methods.
In the interest of backward compatibility, the 486 has a mode bit that says not
to handle FP exceptions automatically, but rather to freeze the FPU and send a
signal on the FERR pin, which is usually tied to an 8259A which then feeds the
interrupt back as IRQ 13. There is some extra complication involved here
because the FPU has to stay frozen until the interrupt is accepted so the CPU
can go back and look at the FPU's state. Early 386/25 chips had a bug that
would sometimes freeze up on a floating point interrupt and you had to get a
kludge socket with a PAL that fixed the timing glitch that provoked the bug.
So as likely as not, the motherboard hardware that runs FERR out and back isn't
working correctly. It's not surprising, few DOS users take floating point
seriously enough to notice whether the interrupts are working right."
When you specify a system, make clear to your vendor that the motherboard must
handle float exceptions properly. Test your motherboard's handling of
divide-by-zero; if it doesn't work, press your vendor to replace it *and
send me email*! Only by publishing a list of boards known bad can we
protect ourselves and pressure vendors to fix this problem.
The 386 UNIX Buyer's Guide posting (pc-unix/software) includes tables of
motherboards and systems known to run with various UNIX ports.
E. Peripherals
Peripherals are another matter, especially hard disks. A good rule of thumb
for balanced configurations is that the hard disk should comprise about half
(or maybe a bit more) of the total system hardware price (exception: if you're
buying a really good monitor, like 16" or over, it's going to be expensive
enough to bust this rule). Unless you're the exception who has to invoke
warranty due to a system arriving dead, most of what you buy from a dealer or
mail-order house is their ability to surf the Winchester market, make volume
buys, and burn in your disks before shipping. We'll look at disk choices in
more detail later on.
You should have a tape drive for backup, and because most UNIX vendors like to
distribute their OS on tape. Ideally, your tape backup should be able to image
your entire disk. Unfortunately, this can get *very* expensive for large
disks; 60/120MB QIC drives are fairly cheap now but larger sizes (typically
150, 250, 525 QIC tapes and 1.3gig DAT) are not (however, if you have large
drives the cost difference can quickly get weaten up by media costs). One
interesting point is that if you've gone SCSI, a 150MB QIC (comparable to the
drives now popular on Suns) may well be cheaper than older 60MB technology due
to differences in the adaptor boards required.
These days, most vendors bundle a 14" monitor and super-VGA card with 1024x768
resolution in with their systems. Details to watch are whether the card comes
loaded with 512K or 1MB of RAM (which will affect how much of that maximum
resolution and how many colors you actually get), whether the memory is
dual-ported VRAM (slightly more expensive but much faster), and whether the
monitor is interlaced or non-interlaced. The latter is better and should no
longer cost extra; look for the abbreviation NI in the ad or quote and be
suspicious if you don't see it.
We'll have much more to say about price/performance tradeoffs in peripherals
in the next major section, on performance tuning.
F. Power Protection
Finally, I strongly recommend that you buy a power conditioner to protect your
hardware. MOV-filtered power bars make nice fuses (they're cheap to replace),
but they're not enough. I've been delighted with my TrippLite 1200, which you
can get for $139 or so by mail order. A fringe benefit of this little beauty
is that if you accidently pull your plug out of the wall you may find you
actually have time to re-connect it before the machine notices!
The tecbical info in the remainder of this section is edited from material
supplied by David E. Wexelblat <d...@mtgzfs3.att.com>.
There are several levels of power protection available to the home computer
user. I break this down into 4 levels; others may have different ways of
classifying things. The levels are:
1. Surge Suppressor
2. Line Conditioners
3. Standby Power Supplies
4. Uninterruptible Power Supplies
and here's what they mean:
1. Surge suppressors
These are basically a fancy fuse between the source and your hardware; they
clamp down spikes, but can't fill in a low voltage level or dropout.
This is a bare minimum level of protection that any piece of expensive
electronics should have. Note that this applies to more than just AC power;
surge suppressors are available for (and should be used on) phone lines, and
RS-232 and parallel connections (for use on long lines; generally not needed if
the devices is colocated with the computer and all devices are protected from
outside sources). Note also that *all* devices connected to your computer need
to be protected; if you put a surge suppressor on your computer but not your
printer, then a zap on the printer may take out the computer, too.
An important fact about surge suppressors is that *they need to be replaced if
they absorb a large surge*. Besides fuses, most suppressors rely on on
components called Metal-Oxide Varistors (or MOVs) for spike suppression, which
degrade when they take a voltage hit. The problem with cheap suppressors is
that they don't tell you when the MOV is cooked, so you can end up with no
spike protection and a false sense of security --- better ones have an
indicator.
You can buy surge suppressors at any Radio Shack; for better prices, go
mail-order through Computer Shopper or some similar magazine. All of
these are low-cost devices ($10-50).
2. Line Conditioners
These devices filter noise out of AC lines. Noise can degrade your power
supply and cause it to fail prematurely. They also protect against short
voltage dropouts and include surge suppression.
My Tripp-Lite 1200 is typical of the better class of line conditioners --- a
box with a good big soft-iron transformer and a couple of moby capacitors in it
and *no* conductive path between the in and out sides. With one of these, you
can laugh at brownouts and electrical storms.
Price vary widely, from $40-400, depending on the power rating and capabilities
of the device. Mail-order from a reputable supply house is your best bet.
Line conditioners typically *don't* need to be replaced after a surge; check
to see if yours includes MOVs.
3. Standby power supplies (SPSs)
These devices are battery-based emergency power supplies that provide power for
your system via an inverter if the power fails. An SPS will generally have all
the capabilities of a line conditioner as well.
Note: these devices do not come on line until after the power fails, and have a
certain amount of delay of some of milliseconds before they come on line. If
the capacitors in your power supply are not large enough, the SPS may not cut
in in time to prevent your computer from seeing the power failure.
Note also that many SPSs are marketed as Uninterruptable Power Supplies (see
below). This is incorrect. Any device with a non-zero cutover time cannot be
a true UPS. If the ad mentions a cutover time, it's an SPS, and not a UPS.
The price range for these devices (depending largely on size and cutover time)
is $200-2000. An SPS will *not* need to be replaced after absorbing a large
surge.
4. Uninterruptable power supplies (UPSs)
These devices provide full-time isolation from the incoming AC line through a
transformer of some sort. These devices are on-line at all times, and if the
AC line fails, the batteries will cut in. Your devices will see no
interruption of their incoming AC. UPSs cost more, and provide more features.
They are the ultimate in power protection. Many UPSs have an intelligent
interface that will notify a connected device of a power failure, allowing it
to shut down cleanly. UPSs also provide the capabilities of a line
conditioner. The price range (for devices in the size range for a home
computer) are $400-$2500. An UPS will *not* need to be replaced after
absorbing a large surge.
Now, given this information, how does one decide what to get? For a system
that runs unattended, like most Unix systems, it is best to have a device that
provides both power holdover and a power failure signal. Hence, for a Unix
system, a UPS is the best idea (an SPS is not the best power protection and
most have no intelligent interface). At least one vendor sells
ordered-shutdown software for Unix, and it's fairly simple to write your own
daemon to monitor a serial port, and send init a SIGPWR signal when it sees a
certain signal.
Our recommendation for a home Unix environment is a configuration like the
following:
a) A true on-line UPS for the computer system. An intelligent
interface is mandatory, along with appropriate software for
ordered shutdown.
b) Surge suppression on all phone lines, and also on serial/parallel
lines that leave the room.
c) Line conditioners on any devices not connected to the UPS. If
you do take a power hit, it's cheaper to replace a $50 line
conditioner than a $1500 laser printer.
An important question is "How do I know how big a UPS to get?" The simple
answer is to add up the VA ratings of all devices you plan to put on it, add
~20%, then add room for growth. For a server-sized system, with a large
monitor, you probably are looking at 800-1100 VA.
One other consideration is that you typically shouldn't put a laser printer on
a UPS. The power draw of the heater will blow the mind of the UPS. David
reports that his UPS (which normally has ~60 minutes of backup) ran out of
juice in 45 seconds after plugging in the laser printer. The other thing is
that you can't even put the laser printer on the same circuit with a UPS ---
the heater kicks on every 20-30 seconds, and most UPSs will see the current
draw as a power failure. So buy a separate line conditioner for the laser
printer.
Finally, read the UPS's installation manual carefully if you're going
to use it with other power-protection devices. Some UPSs don't like having
surge suppressors between them and the equipment.
David personally recommends surge suppressors and line conditioners from
Tripp-Lite (available both mail-order and retail), and UPSs from Best Power
Technologies (Necedah, WI - 1-(800)-356-5737). I can enthusiastically second
the TrippLite recommendation, but haven't dealt with Best Power at all. There
are many other vendors for all of these devices.
Tripp-Lite has a whole range of products, from a $10 phone-line
surge-suppressor, to line conditioners and SPSs with prces in the hundreds of
dollars. They have a line of $50-80 line conditioners that are good for most
peripherals (including your home stereo :->).
Best Power Technologies sells two lines of UPSs in the range for home systems.
The older and more expensive FERRUPS line (which is what David has) has a smart
interface, and very good filtering and surge-suppression capabilities. He says
"I have a 1.15kVA FERRUPS for my home system, which is overkill with my current
hardware (although it rode out a 45 minute power failure with nary a whisper -
no reboot). In 1990, I paid ~$1600 for this device, and that has since gone
up. They also sell a newer line of Fortress UPSs. These are better suited in
price for home systems. I don't know much about them, as they were not
available when I bought my UPS. I expect that this is what most people will
want to consider, though. In addition, Best sells Check-UPS, a software
package (in source form) for monitoring the UPS and shutting it down. I have
found Best to be a good company to deal with, with competent, knowledgable
sales people (who will be able to help you pick the right device), and helpful,
courteous, and responsive technical support."
IV. Performance tuning
Here are the places where you can trade off spending against the performance
level you want to buy and your expected job mix.
A. How to Pick your Processor
Right now, the fastest Intel 386 is 33MHz; AMD makes a 386/40. As we've
discussed, these are minimum for decent UNIX and X performance. To crank
faster than that, you have to go to a 486.
And indeed, the move to 486 processors, especially 486/33DXs, is happening very
fast. On current trends, by the end of '92 it's going to be hard to find an
Intel 386 in any new desktop machine, though the spiffier AMD clone may hang in
there longer (note that the AMD 33DX is as fast as an Intel 40DX, and the AMD
386/40DX seems to scale to sbout the equivalent of a Intel 386/50, if such a
thing actually existed).
Lots of relatively inexpensive 486/33DX systems are out there now; estimates
for the speed advantage from the 486 range from 20 to 50% with 30% pretty
widely heard (thus one of these is equivalent to about a 386/40DX). Most of
the advantage comes from the pipelining and on-chip cache. The higher level of
integration also implies better reliability. And of course, the on-chip FPU
really sizzles if you're into scientific computing or graphics.
Intel's own literature suggests an even heavier advantage for the 486 vs. 386,
at leat 2:1. The following information on this subject appeared in article
<13a29iINN...@iraul1.ira.uka.de> by S_JU...@iravcl.ira.uka.de (|S| Norbert
Juffa):
UNIX performance of Intel processors as given in Intel's literature
Processor SPECmark SPECint SPECfp Whetstone Dhrystone Linpack Ref Rm
double p. 2.1 dp MFLOPS
1) Intel 386/387-33 4.3 6.4 3.3 3290 15888 N/A 1 *+
2) Intel 386/387-33 4.1 6.0 N/A 3200 18900 0.4 2 #
3) RapidCAD-33 6.6 7.3 6.1 5300 18275 N/A 1 *+
4) 486DX-25 8.7 13.3 6.6 5640 32000 1.0 2
5) 486DX-33 11.1 17.5 8.2 7200 43000 1.5 3
6) 486DX-33 12.1 18.3 9.2 N/A N/A N/A 4
7) 486DX-33 14.5 19.0 12.2 12300 43500 1.6 5 &
8) 486DX-50 18.2 27.9 13.6 10710 64400 2.5 3
9) 486DX2-50 19.2 25.4 15.9 18500 63966 2.3 5 &
10)486DX-50 21.9 28.5 18.3 18500 65400 2.4 5 &
11)486DX2-66 25.6 34.0 21.2 24700 85470 3.1 5 &
Remarks:
* Whetstone/Dhrystone are 32-bit DOS results
+ SPEC ratios recomputed from SPEC timings (computed wrong in report)
& note huge increase in SPEC floating point performance over previous results
due to new experimental FORTRAN compiler
# machine with AMD 386-40/Cyrix 83D87-40/128 kB cache is estimated by me at:
7.7 SPECint, 5.0 SPECfp, 6.1 SPECmark,
5600 double prec. Whetstones, 23000 Dhrystones,
0.6 Linpack double prec. MFlops
These estimates based on my own measurements and data from:
FasMath 83D87 Benchmark Report, Cyrix 1990
World's Fastest 386 40 MHz Am386(tm)DX Microprocessor Performance Summary,
AMD 1991
References:
1) Intel RapidCAD(tm) Engineering CoProcessor Performance Brief. 1992
2) i486(tm) Microprocessor Performance Report. 1990.
Order No. 240734-001
3) 50MHz Intel486(tm) DX Microprocessor Performance Brief. 1991.
Order No. 241120-001
4) i486(tm) Microprocessor Business Performance Brief. 1990.
Order No. 281352-002
5) Intel486(tm) DX2 Microprocessor Performance Brief. 1992
Order No. 241254-001
Configurations:
1) COMPAQ SystemPro 386/33 MHz, 8 MB memory, AT&T UNIX System V/386 Release 4.0
Version 2.0
2) 64 kB write back cache,
AT&T UNIX System V Release 3.2CC, MetaWare High C R2.2c,
SVS FORTRAN V2.8
3) COMPAQ SystemPro 386/33 MHz, 8 MB memory, AT&T UNIX System V/386 Release 4.0
Version 2.0
4) 128 kB write-back cache, 12 MB RAM,
AT&T UNIX System V Release 3.2CC, MetaWare High C R2.2c,
SVS FORTRAN V2.8
5) No 2nd level cache, 16 MB RAM,
AT&T UNIX System V/386 R3.2, MetaWare High C R2.3p
SVS FORTRAN V2.8
6) ALR PowerCache 33/4e, 128 kB cache, 16 MB RAM
SCO UNIX System V R3.2.2, MetaWare High C R2.2c/R2.3k,
SVS FORTRAN V 2.8
7) Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
8) 256 kB write-back cache (82495DX/82490DX), 16 MB RAM,
AT&T UNIX System V/386 R3.2, MetaWare High C R2.3p
SVS FORTRAN V2.8
9) Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
10)Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
11)Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
One of Intel's most recent wrinkles is the "clock-doubler" chips. The 50DX2
runs at 25MHz externally but computes at 50MHz. A clock-doubled 33DX (compute
speed 66MHz) is sampling now, and there are persistent rumors of a
clock-doubled 50 in the works that would compute at a blistering 100MHz! Intel
likes to claim a 70% speedup for the doublers over their undoubled brethren.
I've expressed skepticism about this in previous issues, but SPECmarks suggest
that just this once the marketroids may not be lying -- much. Under UNIX, a
50DX2 is in fact nearly as fast as a true 50DX. Still, beware of anyone whose
literature passes off the DX2 qualification in the fine print; they may be
scamming about other things, too.
Right now you'll pay as much as a $1500 premium for a 486/50, as that's
relatively new technology and demands extra-fast memory to run full-out. Also,
these processors run really hot (one correspondent described the 50 as a
"toaster on a chip"). If you go this route, be sure your configuration has an
extra-heavy-duty cooling fan. Or two. And, for preference, a hefty heat sink.
Of course, if you do this you'll be ready to drop in Intel's 50MHz-external/
100MHz-internal part when it comes out later this year, and blow the doors off
all those fancy proprietary-technology workstations.
B. Of Memory In...
Buy lots of RAM, it's the cheapest way to improve real performance on any
virtual-memory system. At $30-$50 maximum per megabyte it's just plain silly
to stick with the 2-4mb now standard on most clone configurations. Go to 8,
you won't regret it; 16 if you're going to use X.
Above 16 is iffy on ISA boxes because the stock USL 4.0.3 kernel may try to do
DMA from a location the bus can't deal with. Most UNIX vendors have fixed this
by adding code that forces DMAs to take place from low memory; make absolutely
sure that includes yours before you load up beyond 16MB. The pc-unix/software
FAQ posting includes information on which vendors are known to have fixed this
problem.
Some motherboards have 16 sockets for SIMM memory modules. Some only 8. Some
take only 1MB mdules, some handle 4MB. These constraints interact in funny
ways.
You should make sure if you are buying an entry level 2 or 6 MB system with a
16-socket motherboard that you will not have to ditch the SIMMs that are
already installed in order to go to your maximum (if 16 MB is your maximum).
Some systems only allow you to mix 1M and 4M SIMMs in certain combinations.
Try not to get any 1M SIMMs in your initial configuration, because you'll
probably end up turfing them later. That is, buy a 4MB, 8MB, 12 MB or 16MB
system to start.
Newer ISA designs have a 32 MB upper limit with only 8 sockets, since they can
take 4Mx9s...however, this means different interleaving (only 2 banks), which
limits the possible configurations. You don't want to start off with an 8 MB
configuration, because that's 8 ea 1Mx9's, filling up all the sockets...the
next upgrade requires replacing 1Mx9 with 4Mx9. You can't even set up 12
MB!...the first reasonable config (that won't require tossing hardware) is 16
MB, since that's one bank full of 4Mx9.
Most EISA motherboards have 16 4MB-capable sockets, and this is clearly
where the market is going.
C. Bus wars
Should you buy 16-bit ISA vs. 32-bit EISA? You'll pay a $600-$900 premium for
the latter. What you get in return is the ability to use things like fast
32-bit SCSI controllers and a smoother upward-migration path. On the other
hand, EISA cards are significantly more expensive. And so far, there isn't
much support for EISA-specific hardware --- a couple of vendors will drive EISA
SCSI disk and tape controllers and that's about it (of course those *are* the
most important bandwidth-eaters). All ISA cards will still work.
Of course, most of what you get from EISA is a performance boost. There are
two different theories about why EISA is better; both have their adherents.
Theory A: Bandwidth matters
UNIX has always been an I/O-intensive operating system. According to this
theory, increasing processor speed on clones can leave it spending all its time
waiting on the limited I/O capacity of the poor old 5.3MB/sec ISA bus. The
vendors all seem to think this starts at around 33MHz and that if you're buying
50MHz it definitely pays to go EISA.
Theory B: Cache is what matters
According to this theory, UNIX never comes even close to saturating the ISA-bus
bandwidth. EISA boards are faster because the premium vendors can charge for
them allows the motherboard designer more freedom and a richer parts budget.
The most important performance effect of this is that EISA boards have larger
and better-designed caches, increasing the effective memory-access speed.
There's probably some truth to both analyses. If your machine is going to
spend most of its processor time running X displays and doing other classically
compute-bound tasks, cache size matters most. On the other hand, benchmarks
show that the combination of TCP/IP and multi-user disk activity *can* saturate
ISA, and one can sometimes *see* a fast-processor machine slow down during disk
accesses...
If you're contemplating any kind of heavy-duty networking, EISA network
adapters will become rather important. A correspondent tells me he's seen
benchmarks showing what percentage of bus bandwidth is consumed by various
cards when flooding an ethernet (i.e. consuming the entire 10Mbit bandwidth of
a quiet net, as you might be when doing an FTP transfer, for instance). 8-bit
ISA cards consume 40-60% of bus bandwidth; 16-bit cards, 20-40%. 32-bit EISA
cards consume only about 5-10%. This would be particularly important in a
machine being used as a bridge, where you might be handling a large portion of
the traffic on two or more separate nets. The advantage of EISA cards may be
due to their shorter-cycle bus mastering DMA. At time of writing, only
SCO supports these, but other UNIX vendors are known to have their own drivers
in the pipeline.
D. IDE vs. SCSI
Another basic decision is IDE vs. SCSI. Either kind of disk costs about the
same, but the premium for a SCSI card varies all over the lot, partly because
of price differences between ISA and EISA SCSI cards and especially because
many motherboard vendors bundle an IDE chip right on the system board. SCSI
gives you better speed and throughput and loads the processor less, a win for
larger disks and an especially significant consideration in a multi-user
environment; also it's more expandable.
Another important win for SCSI is that it handles multiple devices much more
efficiently. If you have two IDE (or ST506 or ESDI) drives, only one can
transfer between memory and disk at once. In fact, you have to program them at
such a low level, one drive might actually be blocked from *seeking* while
you're talking to the other drive. SCSI drives are mostly autonomous and can
do everything at once; and current SCSI drives are not quite fast enough to
flood more than 1/2 the SCSI bus bandwidth, so you can have at least two drives
on a single bus pumping full speed without using it up. In reality, you don't
keep drives running full speed all the time, so you should be able to have 3-4
drives on a bus before you really start feeling bandwidth crunch.
All this having been said, don't write off IDE too quickly. Sure, it's
compatible with the nasty old ST506 interface, but it's *much* faster. It
remains the cost-effective choice for smaller drives (up to 500MB) on systems
that won't be hitting the disk constantly. Unless you're running a heavily
used network or database server, don't assume SCSI will make any noticeable
difference.
(If you can do your own installation, I hear that used 150/250MB SCSI drives
are getting quite common and cheap on the net. All 150MB QIC type drives can
do 250MB on extended-length tapes, though some manufacturers discourage you
from doing this to avoid excessive heade wear. But back to disks...)
The following, by Ashok Singhal <ash...@duckjibe.eng.sun.com> of Sun
Microsystems, is a valiant attempt to demystify SCSI terminology.
The terms "SCSI" and "SCSI-2" refer to two different specifications.
Each specification has a number of options. Many of these options are
*independent* of each other. I like to think of the main options (there are
others that I'll skip over because I don't know enough about them to talk
about them on the net) by classifying them into five categories:
1. Logical
This refers to the commands that the controllers understand.
SCSI-2 defined a common cammand set that is pretty much a
superset of the SCSI command set.
2. Data Width
8 bits (+ 1 parity) -> "normal"
16-bits (+ 2 parity) -> "wide"
32-bits (+ 4 parity) -> I don't know, "extra-wide??"
All three options are available in SCSI-2 (yes,
the draft spec I have even shows 32-bits!), although
8-bit wide is still by far the most common. Not sure, but I believe
SCSI defined only 8-bit wide data path.
3. Electrical Interface
single-ended (max cable length 6 meters)
differential (max cable length 25 meters)
Both options are available for SCSI-2 (I'm not sure about SCSI,
but I think both options were available also)
and this option is independent of options 2, 4, 5. Differential
is less common but allows better noise immunity and longer
cables.
4. Handshake
Asynchronous (requests and acks alternate)
Synchronous (multiple requests can be outstanding)
Both options are available for SCSI-2 (Not sure about SCSI,
but I think both were available also). This is negotiated
between each target and initiator; asynchronous and synchronous
transfers can occur on the same bus. This is independent of
2, 3 (Not sure about 1).
5. Synchronous Speed (does not apply for asynchronous option)
"Normal" is up to 5 Mtransfers/sec ( = 5MB/s for 8-bit wide, more
for wider)
"Fast" is up to 10 Mtransfers/s ( = 10 MB/s for 8-bit wide, more
for wider)
The fast option is defined only in SCSI-2.
This options basically defines shorter timing parameters
such as the assertion period and hold time.
The parameters of the synchronous transfer are negotiated
between each target and initiator so different speed transfers
can occur over the same bus.
E. Other Disk Decisions
Look at seek times and transfer rates for your disk; under UNIX disk speed and
throughput are so important that a 1-millisecond difference in average seek
time can be noticeable.
Previous issues said "Disk caching is good, but there can be too much of a
good thing. Excessively large caches will slow the system because the overhead
for cache fills swamps the real accesses (this is especially a trap for
databases and other applications that do non-sequential I/O). More than 100K
of cache is probably a bad idea for a general-purpose UNIX box; watch out for
manufacturers who inflate cache size because memory is cheap and they think
customers will be impressed by big numbers." This may no longer be true on
current hardware; in particular, most controllers will interrupt a cache-fill
to fulfill a `real' read request.
In any case, having a large cached hard drive (particularly in the IDEs) often
does not translate to better performance. For example, Quantum makes a 210Mb
IDE drive which comes with 256Kb cache. Conner and Maxtor also have 210Mb
drives, but only with 64Kb caches. The transfer rate on the drives, however,
show that the Quantum comes in at 890Kb/sec, while the Maxtor and Conner fly
away at 1200Kb/sec. Clearly, the Conner and Maxtor make much better use of
their smaller cache.
Many retailers seem to enjoy advertising the "9ms" Quantum 52/80/120/200Mb
drives. This speed, of course, is bogus. All the quantum drives are at least
16ms is average access. The 9ms already includes the cacheing speedup.
However, it may be that *any* hardware disk caching is a lose for UNIX! Scott
Bennett <benn...@mp.cs.niu.edu> reports a discussion on comp.unix.wizards:
"nobody found the hardware disk caches to be as effective in terms of
performance as the file system buffer cache...In many cases, disabling the
hardware cache improved system performance substantially. The interpretation
of these results was that the caching algorithm in the kernel was superior to,
or at least better tuned to UNIX accesses than, the hardware caching
algorithms."
Thus, if your disk controller allows it, try disabling the cache. Your
throughput may go up!
F. Souping Up X Performance
One good way to boost your X performance is to invest in a graphics card with a
dedicated blitter or high-speed local-bus connection, like the ATI series or
the S3-based Quantum, Wind/X and Orchid Fahrenheit 1280. A number of clone
vendors offer these accelerator options relatively cheap and can make your X go
like a banshee; however, stock X doesn't support them yet. Third-party servers
are available from MetroLink (email sa...@metrolink.com) or SGCS
(i...@sgcs.com).
Here is a current price list from MetroLink:
Runtime (all servers, standard and contrib clients) : 299.00
Development (full X11 and Motif 1.1.4 libraries) : 299.00
Xv - Real-Time Video in an X window (true server : 99.00
extension)
Xie - X Imaging Extension : 199.00
And here is the corresponding info from SGCS:
Ref # Description Price
----- --------------------------------------------- ------
** 1 Full X11R5 binaries licensed for a single CPU 295.00
2 Additional X11R5 binary license 50.00
** 3 Enhanced X11R5 source code (note 4) 195.00
4 MIT source code of contributed clients (note 5) 50.00
** 5 Motif binaries for a single CPU 245.00
6 Additional Motif binary license 60.00
** 7 X11R5 Documentation Set 150.00
8 PHIGS Documentation Set 75.00
** DISCOUNTS:
If your choose more than one selection from any of the (**) items above
you will receive the following discounts: $50 off on 2 selections,
$75 off on 3 selections, $100 off on 4 selections
S3 is an interesting phenomenon. Though several vendors advertise their S3
connector as "proprietary" it's actually an anticipation of a forthcoming VESA
(Video Electronics Standards Association) standard. S3 cards are thus
semi-interchangeable.
In general, the ATI approach (normal bus, dedicated blitter and optimization
for special functions like character drawing) will speed up text display, text
scrolling and window resize/move operations a lot, but line-drawing and
graphics only a little. S3, on the other hand, speeds up high-bandwidth
graphics drawing a lot but doesn't have as big an advantage for ordinary
text operations. You pays your money and takes your choice. Benchmarks
indicate that most non-CAD users are better served by the ATI approach.
If you're feeling *really* flush, plump for a 15", 17" or even 20" monitor.
The larger size can make a major difference in viewing comfort. Also you'll be
set for VESA 1280x1024 when everybody gets to supporting that. In the mean
time, the bigger screen will allow you to use fonts in smaller pixel sizes so
that your text windows can be larger, giving you a substantial part of the
benefit you'd get from higher pixel resolutions.
V. Of Mice and Machines
In a previous issue, I claimed that all mice and trackballs are the same for
compatibility purposes. I was wrong -- seriously wrong. The more I found out,
the messier the picture gets. The following is an attempt to sort out all the
confusion. Thanks to Jim McCarthy at Logitech for digging into the matter
and somewhat alleviating my ignorance.
Mice and trackballs used to be simple; now, thanks to Microsoft, they're
complicated. In the beginning, there was only the Mouse Systems 3-button
serial mouse; this reported status to a serial port 30 times a second using a
5-byte serial packet encoding now called "C" protocol. The Logitech Series 7
and 9 mice were Mouse Systems-compatible. All UNIXes that have any mouse
support at all understand C-protocol serial mice.
Then Microsoft got into the act. They designed a two-button serial mouse which
reports only deltas in a three-byte packet; that is, it sends changes in button
status and motion reports only when the mouse is actually moving. This is
called `M' protocol. Microsoft sold a lot of mice, so Logitech switched from
`C' to `M' --- but they added a third button, state changes for which show up
in an optional fourth byte. Thus, `M+' protocol, upward-compatible with
Microsoft's `M'. Most UNIX vendors add support for M+ mice, but it's wise to
check.
Bus mice are divided into 8255 and InPort types. These report info
continuously at 30 or 60 Hz (though InPort mice have an option for reporting
deltas only), and you get interrupts on events and then have to poll hardware
ports for details. More on these next issue.
In addition to serial mice and bus mice, there are "keyboard mice". On PS/2s
there are two identical-looking keyboard ports, labeled (with icons) "mouse" &
"keyboard". Both are 8 or 9 pin mini-DIN's that look like the regular PC
keyboard port only smaller. I don't know what logical protocol the keyboard
mouse speaks. Physically, the connector is eventually connected to the
keyboard processor (often an 8042). The same keyboard processor that decodes
the keyboard decodes the mouse. PS/2s have this port, many newer ISA/EISA
motherboards do as well.
All things considered, UNIX users are probably best off going with a serial
mouse (most current clone motherbords give you two serial ports, so you can
dedicate one to this and still have one for the all-important modem). Not only
are the compatibility issues less daunting, but a serial mouse loads the
multitasking system less due to interrupt frequency. Beware that most clone
vendors, being DOS oriented, bundle M-type mice for which UNIX support is
presently spotty, and they may not work with your X. Ignore the adspeak about
dpi and pick a mouse/trackball that feels good to your hand.
VI. When, Where and How to Buy
If you're a serious UNIX hacker for either fun or profit, you're probably in
the market for what the mail-order vendors think of as a high-end or even
`server' configuration, and you're going to pay a bit more than the DOS
lemmings. On the other hand, prices keep dropping, so there's a temptation to
wait indefinately to buy. A tactic that makes a lot of sense in this market,
if you have the leisure, is to fix in your mind a configuration and a trigger
price that's just a little sweeter than the market now offers and buy when
that's reached.
Direct-mail buying makes a lot of sense today for anyone with more technical
savvy than J. Random Luser in a suit. Even from no-name mail-order houses,
parts and system quality tend to be high and consistent, so conventional
dealerships don't really have much more to offer than a warm fuzzy feeling.
Furthermore, competition has become so intense that even mail-order vendors
today have to offer not just lower prices than ever before but warranty and
support policies of a depth that would have seemed incredible a few years back.
For example, many bundle a year of on-site hardware support with their medium-
and high-end "business" configurations for a very low premium over the bare
hardware.
Note, however, that assembling a system yourself out of mail-order parts is
*not* likely to save you money over dealing with the mail-order systems
houses. You can't buy parts at the volume they do; the discounts they command
are bigger than the premiums reflected in their prices. The lack of any
system warranty or support can also be a problem even if you're expert enough
to do the integration yourself --- because you also assume all the risk of
defective parts and integration problems.
Cruise through "Computer Shopper" and similar monthly ad compendia. Even if
you decide to go with a conventional dealer, this will tell you what *their*
premiums look like.
Another alternative to conventional dealerships (with their designer "looks",
stone-ignorant sales staff, and high overheads that *you* pay for) is to go
with one of the thousands of the hole-in-the-wall stores run by immigrants from
the other side of the International Date Line. They're usually less ignorant
and have much lower overheads; they do for you locally what a mail-order house
would, that is assemble and test parts they get for you from another tier of
suppliers. You won't get plush carpeting or a firm handshake from a white guy
with too many teeth and an expensive watch, but then you didn't really want to
pay for those anyway, right?
A lot of vendors bundle DOS 5.0 and variable amounts of DOS apps with their
hardware. You can tell them to lose all this cruft and they'll shave $50 or
$100 off the system price. However, David Wexelblat observes "there are at
least two situations in which the Unix user will need DOS available: 1) most,
if not all, EISA configuration utilities run under DOS, and 2) SCSICNTL.EXE by
Roy Neese is a godsend for dealing with SCSI devices on Adaptec boards."
Don't forget that (most places) you can avoid sales tax by buying from an
out-of-state mail-order outfit, and save yourself 6-8% depending on where you
live. If you live near a state line, buying from a local outfit you can often
win, quite legally, by having the stuff shipped to a friend or relative just
over it. Best of all is a buddy with a state-registered dealer number; these
aren't very hard to get and confer not just exemption from sales tax but
(often) whopping discounts from the vendors. Hand him a dollar afterwards to
make it legal.
(Note: I have been advised that you shouldn't try the latter tactic in
Florida -- they are notoriously tough on "resale license" holders).
(Note II: The Supreme Court recently ruled that states may not tax out-of-state
businesses under existing law, but left the way open for Congress to pass
enabling legislation. Let's hope the mail-order industry has good lobbyists.)
On the other hand, one good argument for buying locally is that you may have to
pay return postage if you ship the system back. On a big, heavy system, this
can make up the difference from the savings on sales tax.
VII. Questions You Should Always Ask Your Vendor
A. Minimum Warranty Provisions
The weakest guarantee you should settle for in the mail-order market should
include:
* 72-hour burn-in to avoid that sudden infant death syndrome. (Also,
try to find out if they do a power-cycling test and how many repeats
they do; this stresses the hardware much more than steady burn-in.)
* 30 day money-back guarantee. Watch out for fine print that weakens this
with a restocking fee or limits it with exclusions.
* 1 year parts and labor guarantee (some vendors give 2 years).
* 1 year of 800 number tech support (many vendors give lifetime support).
Additionally, many vendors offer a year of on-site service free. You should
find out who they contract the service to. Also be sure the free service
coverage area includes your site; some unscrupulous vendors weasel their way
out with "some locations pay extra", which translates roughly to "through the
nose if you're further away than our parking lot".
If you're buying store-front, find out what they'll guarantee beyond the
above. If the answer is "nothing", go somewhere else.
B. Documentation
Ask your potential suppliers what kind and volume of documentation they supply
with your hardware. You should get, at minimum, operations manuals for the
motherboard and each card or peripheral; also an IRQ list, and a bad-block
listing if your Winchester is ESDI rather than IDE or SCSI (the latter two
types of drive do their own bad-block mapping internally). Skimpiness in this
area is a valuable clue that they may be using no-name parts from Upper
Baluchistan, which is not necessarily a red flag in itself but should prompt
you to ask more questions.
C. A System Quality Checklist
There are various cost-cutting tactics a vendor can use which bring down the
system's overall quality. Here are some good questions to ask:
* Is the memory zero-wait-state? One or more wait states allows the vendor to
use slower and cheaper memory but will slow down your actual memory subsystem
throughput. This is a particularly important question for the *cache*
memory!
* Is the monitor non-interlaced? Does it have a tilt-and-swivel base? Is it
*color*? Yes, if you don't see it in the ad, ask; some lowball outfits will
try to palm off so-called "black & white VGA" monitors on you. What's the
vertical scan rate? 60Hz is SVGA standard; 72Hz is VESA standard and
minimal for flicker-free operation; 80Hz is cutting-edge. What's the dot
pitch? .31mm is minimal, .28mm or .27mm is good. You need .28mm for X.
A slightly larger dot pitch is acceptable in a larger monitor (16" or more).
* If you're buying a factory-configured system, does it have FCC certification?
While it's not necessarily the case that a non-certified system is going
to spew a lot of radio-frequency interference, certification is legally
required --- and becoming more important as clock frequencies climb. Lack
of that sticker may indicate a fly-by-night vendor, or at least one in
danger of being raided and shut down!
VIII. Things to Check when Buying Mail-Order
A. Tricks and Traps in Mail-Order Warranties
Reading mail-order warranties is an art in itself. A few tips:
Beware the deadly modifier "manufacturer's" on a warranty; this means you have
to go back to the equipment's original manufacturer in case of problems and
can't get satisfaction from the mail-order house. Also, manufacturer's
warranties run from the date *they* ship; by the time the mail-order house
assembles and ships your system, it may have run out!
Watch for the equally deadly "We do not guarantee compatibility". This gotcha
on a component vendor's ad means you may not be able to return, say, a video
card that fails to work with your motherboard.
Another dangerous phrase is "We reserve the right to substitute equivalent
items". This means that instead of getting the high-quality name-brand parts
advertised in the configuration you just ordered, you may get those no-name
parts from Upper Baluchistan --- theoretically equivalent according to the
spec sheets, but perhaps more likely to die the day after the warranty expires.
Substitution can be interpreted as "bait and switch", so most vendors are
scared of getting called on this. Very few will hold their position if you
press the matter.
Another red flag: "Only warranted in supported environments". This may mean
they won't honor a warranty on a non-DOS system at all, or it may mean they'll
insist on installing the UNIX on disk themselves.
One absolute show-stopper is the phrase "All sales are final". This means you
have *no* options if a part doesn't work. Avoid any company with this policy.
B. Special Questions to Ask Mail-Order Vendors Before Buying
* Does the vendor pay for shipping? What's the delivery wait?
* If you need to return your system, is there a restocking fee? and will the
vendor cover the return freight? Knowing the restocking fee can be
particularly important, as they make keep you from getting real satisfaction
on a bad major part. Avoid dealing with anyone who quotes more than a 15%
restocking fee --- and it's a good idea, if possible, to avoid any dealer
who charges a restocking fee at all.
C. Payment Method
It's a good idea to pay with AmEx or Visa or MasterCard; that way you can
stop payment if you get a lemon, and may benefit from a buyer-protection plan
using the credit card company's clout. However, watch for phrases like "Credit
card surcharges apply" or "All prices reflect 3% cash discount" which mean
you're going to get socked extra if you pay by card.
VIII. Which Clone Vendors to Talk To
I went through the March 1992 issue of Computer Shopper calling vendor 800
numbers with the following question: "Does your company have any
configurations aimed at the UNIX market; do you use UNIX in-house; do
you know of any of the current 386 or 486 ports running successfully
on your hardware?
I didn't call vendors who didn't advertise an 800 number. This was only partly
to avoid phone-bill hell; I figured that toll-free order & info numbers are so
standard in this industry sector that any outfit unable or unwilling to spring
for one probably couldn't meet the rest of the ante either. I also omitted
parts houses with token systems offerings and anybody who wasn't selling
desktops or towers with a 386/33DX or heavier processor inside.
After plundering Computer Shopper, I called up a couple of "name" outfits that
don't work direct-mail and got the same info from them.
The answers I get revealed that for most clone vendors UNIX is barely a blip on
the screen. Only a few have tested with an SVr4 port. Most seem barely aware
that the market exists. Many seem to rely on their motherboard vendors to tell
them what they're compatible, without actually testing whole systems. Since
most compatibility problems have to do with peripheral cards, this is a
problem.
Here's a summary of the most positive responses I got:
A --- Advertises UNIX compatibility.
C --- Has known UNIX customers.
I --- Uses UNIX in-house.
T --- Have formally tested UNIX versions on their hardware.
F --- Have 486/50 systems
* --- Sounded to me like they might actually have a clue about the UNIX market.
Vendor A C I T F * Ports known to work
--------------- - - - - - - -----------------------------------------------
ARC . . X X . . SCO XENIX 2.3.2, SCO UNIX 3.2.1
AST . X X X X * SCO UNIX 3.2.4, ODT 2.0 Microport V/4
Allegro . . X X . . SCO XENIX 3.2.4
Altec . X . X . . XENIX (no version given).
Ares . X X X X * AT&T 3.2, ISC (version unknown)
Basic Time . X X X X * SCO XENIX 2.3.2, have in-house UNIX experts.
Binary Tech . X . X X . Claims to work with all versions.
Blue Dolphin . X . X X * SCO XENIX.
CCSI X X . . X . They've used SCO XENIX, no version given.
CIN . X . . . . SCO UNIX (version not specified)
CSS . X . X . * SCO 3.2.2, ISC 3.0, SCO ODT. See Will Harper.
Centrix X . . . . . No specifics on versions.
Compudyne . X X X X . Couldn't get details on which versions.
Comtrade . X . X X . Couldn't get details on which versions.
Datom X X X X X . SCO XENIX 3.2.
Dell X X X X X * See Dell SVr4 data.
Desert Sands X X . X X . SCO UNIX 3.2.4
Digitech . X . X . . SCO UNIX 3.2.1, XENIX 2.3.1
EPS X X X X . * SCO XENIX 3.2.4, ISC & AT&T (versions not sp.)
Gateway 2000 X X X X X * SCO UNIX 3.2.0. XENIX 2.3.4 ISC 3.0, ESIX 4.0.3
HD Computer . X . X X . SCO UNIX 3.2, SCO XENIX 3.2.2
HiQ . X . X . . SCO UNIX (version not specified)
Infiniti . X . X X . SCO UNIX (versions not specified)
Insight . . X . X . SCO XENIX 3.2.4. No tech support for UNIX
Keydata X . X X X * SCO version 4, ISC 3.2
Legatech . X . . X . SCO UNIX, ISC (versions not specified)
MicroGeneration . . X . . . Uses XENIX.
MicroLab X . . . . . SCO UNIX, SCO XENIX
MicroSmart X X . X . . SCO XENIX (version not specified)
Microlink X . . X X . SCO XENIX (version not specified)
Myoda X X . X X . SCO XENIX 3.2.2, ISC 3.2
Naga . X . X X * SCO & XENIX 3.2.
Northgate X X . X X * SCO UNIX 3.2
PC Brand . X X X . . SCO XENIX, ISC UNIX
PC Professional . X . X . . ISC 3.2
PC-USA X X . X . . ISC 5.3.2 and SCO 3.2
Profex . X . X . . SCO XENIX 3.2.
Royal Computer . X . . X . No details on versions.
SAI X X . X X . SCO UNIX 3.2.2.
Santronics . . X X X . SCO XENIX 3.2.4
Solidtech . X . . . . Dell (no version given), ISC 3.2.
Strobe . . . X X . SCO, Microport, ISC (no version numbers given)
Swan X X X X X * SCO 2.3.1, UNIX 3.2, ISC 3.2v2.0.2
TriStar . X X X X * SCO UNIX 3.2.2, XENIX 2.3.2, ISCr4
Zenon . X . X X * SCO UNIX (version not specified)
Zeos . X X X X * SCO XENIX 3.2.4, AT&T 3.2
Special notes about a few vendors who appear to have a clue:
Ares targets some of its systems for UNIX CAD use. They have a house wizard
name Ken Cooper (everybody calls him "K.C.").
EPS targets some 486 EISA configurations for UNIX.
Swan doesn't know the UNIX market very well yet, but their project manager
wants a bigger piece of it and is interested in doing some of the right
things. They have a house wizard, one John Buckwalter.
Dell, of course, supports an industry-leading SVr4 port. They're a bit on
the pricy side, but high quality and very reliable. Lots of UNIX expertise
there; some of it hangs out on the net.
Zeos is on the net as zeos.com, with a uunet connection; they host a UNIX BBS.
They have an in-house UNIX group reachable at supp...@zeos.com; talk to Ken
Germann for details. There are biz.zeos.general and biz.zeos.announce groups
on USENET.
Special notes about a lot of vendors who appear to have *no* clue:
Vendors where I couldn't get a real person on the line, either because
no one answered the main number or because I couldn't raise anyone at
tech support after being directed there: Sunnytech, Quantex, AMS, USA
Flex, Lapine, Syntax Computer, MicroTough, PAC International, The Portable
Warehouse.
Vendors where the question met with blank incomprehension, puzzlement,
consternation, or "We've never tested with UNIX": Allur, AmtA, Aplus, HiTech,
Locus Digital Products, LodeStar, TriStar Computers, Ultra-Comp, UTI Computers,
PC Turbo Corp, Evertek, Microcomputer Concepts, Jinco Computers, UWE, ToughCom,
System Dynamics Group, Terribly Fast Bus Systems.
Vendors who understood the questions but had no answer: Bulldog Computer
Products, LT Plus, Standard Computer, JCC.
Vendors who said "Yes, we're UNIX-compatible" but had no details of any tests:
CompuCity.
Vendors who said "Go ask our motherboard vendor": Ariel Design, Lucky Computer
Co., V-com, Professional Computer, MicroLine, MileHi.
Vendors who sent me to a toll number: Absec, Hokkins, New Technologies, Mirage.
Vendors that believe they have UNIX customers, but can't be any definite than
that: Austin Computer Systems, PC Professional, Treasure Chest Computer
Systems, CompuAdd Express, FastMicro, MidWest Micro.
Final note:
If you order from these guys, be sure to tell them you're a UNIX customer
and don't need the bundled DOS. This will shave some bucks off the system
price, *and* it may encourage them to pay more attention to the UNIX market.
--
Send your feedback to: Eric Raymond = e...@snark.thyrsus.com
Path: sparky!uunet!ornl!sunova!linac!pacific.mps.ohio-state.edu!cis.ohio-state.edu!
rutgers!cbmvax!snark!esr
From: e...@snark.thyrsus.com (Eric S. Raymond)
Newsgroups: comp.unix.sysv386,comp.sys.intel,news.answers
Subject: PC-Clone UNIX Hardware Buyer's Guide
Summary: Tips on how and where to buy hardware for your UNIX.
Message-ID: <1jWcGn#4fvVtm5Qb2WK1ZYkQn0Nhlpc=esr@snark.thyrsus.com>
Date: 20 Nov 92 21:42:27 GMT
Expires: 20 Feb 93 00:00:00 GMT
Sender: e...@snark.thyrsus.com (Eric S. Raymond)
Followup-To: comp.unix.sysv386
Lines: 1429
Approved: news-answers-requ...@MIT.Edu
Archive-name: pc-unix/hardware
Last-update: Fri Nov 20 16:16:32 1992
Version: 8.0
[Note: This issue is being released nearly two weeks ahead of schedule. This
is because (a) there have been been big changes in the hardware market since
7.0 --- the 386 is dead, long live the 386, and (b) I received an unusually
large number of additions and corrections during the last week, including
material for a new major section.]
What's new in this issue:
* The 386 is history.
* Advantages of PS/2 mice.
* A new major section on tape drive types and formats.
* Info on the ClariNet Street Price Report
Gentle Reader: if you end up buying something based on information from this
Guide, please do yourself and the net a favor; make a point of telling the
vendor "Eric's FAQ sent me" or some equivalent. The idea isn't to hype me
personally, I've already got all the notoriety I need from doing things like
_The_New_Hacker's_Dictionary_ --- but if we can show vendors that the Guide
influences a lot of purchasing decisions, I can be a more powerful advocate for
the net's interests, and for you.
0. Contents
I. Introduction
II. Overview of the Market
III. Buying the basics
A. Getting Down to Cases
B. Power Supplies and Fans
C. The Heart Of The Machine
D. Motherboards and BIOSes
E. Peripherals
F. Keyboards
G. Power Protection
IV. Performance tuning
A. How To Pick Your Processor
B. Of Memory In...
C. Bus Wars
D. IDE vs. SCSI
E. Other Disk Decisions
F. Souping Up X Performance
V. Tape Drive Follies
VI. Of Mice and Machines
VII. When, Where and How to Buy
VIII. Questions You Should Always Ask Your Vendor
A. Minimum Warranty Provisions
B. Documentation
C. A System Quality Checklist
IX. Things to Check when Buying Mail-Order
A. Tricks and Traps in Mail-Order Warranties
B. Special Questions to Ask Mail-Order Vendors Before Buying
C. Payment Method
X. Which Clone Vendors to Talk To
I. Introduction
The purpose of this posting is to give you the background information you need
to be a savvy buyer of 386/486 hardware for running UNIX. It is aimed
especially at hackers and others with the technical skills and confidence to go
to the mail-order channel, but contains plenty of useful advice for people
buying store-front retail. It was formerly part of 386-buyers-faq issues 1.0
through 4.0, and is still best read in conjunction with the pc-unix/software
FAQ descended from that posting.
This document is maintained and periodically updated as a service to the net by
Eric S. Raymond <e...@snark.thyrsus.com>, who began it for the very best
self-interested reason that he was in the market and didn't believe in plonking
down several grand without doing his homework first (no, I don't get paid for
this, though I have had a bunch of free software and hardware dumped on me as a
result of it!). Corrections, updates, and all pertinent information are
welcomed at that address. The editorial `we' reflects the generous
contributions of many savvy USENETters.
This posting is periodically broadcast to the USENET group comp.unix.sysv386
and to a list of vendor addresses. If you are a vendor representative, please
check to make sure any information pertaining your company is current and
correct. If it is not, please email me a correction ASAP. If you are a
hardware-knowledgeable user, please send me any distillation of your
experience that you think might improve this posting.
II. Overview of the Market
The central fact about 386/486 clone hardware that conditions every aspect of
buying it is this: more than anywhere else in the industry, de-facto hardware
standards have created a commodity market with low entry barriers, lots of
competitive pressure, and volume high enough to amortize a *lot* of development
on the cheap.
The result is that this hardware gives you lots of bang-per-buck, and it's
getting both cheaper and better all the time. Furthermore, margins are thin
enough that vendors have to be lean, hungry, and *very* responsive to the
market to survive. You can take advantage of this, but it does mean that much
of the info in the rest of this document will be stale in three months and
completely obsolete in six.
One good general piece of advice is that you should avoid both the highest-end
new-technology systems (those not yet shipping in volume) and the very cheapest
systems put out by vendors competing primarily on price. The problem with
the high end is that it usually carries a hefty "prestige" price premium, and
may be a bit less reliable on average because the technology hasn't been
through a lot of test/improve cycles. The problem with the low end is that
price-cutters sometimes settle for marginal components. UNIX is more
sensitive to hardware flakiness than DOS, so cut-price systems that might
deliver consistently for DOS lemmings can come around and bite you. Use a
little care, and spend the $200-$300 to stay out of the basement. It's worth
it.
The last point deserves a little amplification. In the PC world, there's a lot
of "if it doesn't fail, it's OK". It is common to ignore normal engineering
tolerances --- the allowances for variations in components, temperature,
voltage margins, and the like --- and to assume that anything which doesn't
fail outright must work. Watch out! For example, The ISA bus was originally
designed for 6 MHz. IBM later updated that to 8 MHz, and that's as much of a
standard as there is, yet there are motherboards that will let you (try to!)
run it at 12 MHz --- 50% over spec. Some cards are actually designed to work
at that speed with proper tolerances. Others might work...or they might flake
out when they get warm. Any systems vendor above the fly-by-night level is
going to shoot for a little more reliability than this, burning in systems and
(often) doing at least a token system test with some kind of UNIX (usually
XENIX). Pay the few extra bucks it costs to deal with a more careful vendor.
Memory sufficiently fast and reliable for 486/50DX systems running UNIX seems
to be a particular problem (anything slower than 60ns will hurt performance).
The following war story by one comp.unix sysv386 regular is typical: "Dell 2.2
ran perfectly on 3 different AMI 486/50 EISA boards. That is, after I replaced
faulty memory chips which caused repeated panics. My conclusion, after
consulting with our hardware suppliers, was that current quality control on
top-end memory chips (NEC, Toshiba) is not good enough for 486/50's running
serious Unix. The memory will pass every DOS-based test. One has to plug and
play to get a set of simms that work reliably. Part of the hazerds of leading
edge technology."
In mid-November, one correspondent recommended Goldstar Gold-on-Gold 1x3 or
1x9. The idiots in Congress (a redundant phrase if ever there was one) have
imposed an "anti-dumping" (read: "protect American fat-cats") tariff that
immediately jacked up prices by $20 per megabyte. Just in time for Xmas...
III. Buying the basics
In this section, we cover things to look out for that are more or less
independent of price-performance tradeoffs, part of your minimum system
for running UNIX.
A. Getting Down to Cases
Cases are just bent metal. It doesn't much matter who makes those, as long as
they're above an easy minimum quality (on some *really* cheap ones, cards fail
to line up nicely with the slots, drive bays don't align with the access
cutouts, or the motherboard is ill-supported and can ground out against the
chassis). If you're fussy about RFI (Radio-Frequency Interference), it's worth
finding out whether the plastic parts of the case have conductive coating on
the inside; that will cut down emissions significantly, but a few cheap cases
omit it.
Should you buy a desktop or tower case? Our advice is go with tower unless
you're building a no-expansions personal system and expect to be using the
floppies a lot. Many vendors charge nothing extra for a tower case and the
absolute *maximum* premium I've seen is $100. What you get for that is less
desktop clutter, more and bigger bays for expansion, and often (perhaps most
importantly) a beefed-up power-supply and fan. Putting the box and its fan
under a table is good for maybe 5db off the effective noise level, too.
Airflow is also an issue; if the peripheral bays are less cramped, you get
better cooling. This is a good argument for a full tower rather than the `baby
tower' cases some vendors offer. Be prepared to buy extension cables for your
keyboard and monitor, though; vendors almost never include enough flex.
For users with really heavy expoansibility requirements, rackmount PC cases do
exist (ask prospective vendors). Typically a rackmount case will have pretty
much the same functionality as an ordinary PC case. But, you can then buy
drive racks (complete with pwer supply), etc. to expand into. Also, you can
buy passive backplanes with up to 20 or so slots. You can either put a CPU card
in one of the slots, or connect it to an ordinary motherboard through one of
the slots.
B. Power Supplies and Fans
Power supplies can matter but quality is cheap; give preference to those with a
Underwriter's Laboratories rating. There's some controversy over optimum
wattage level; on the one hand, you want enough wattage for expansion. On the
other, big supplies are noisier, and if you draw *too little* current for the
rating the delivered voltage can become unstable. And the expected wattage
load from peripherals is droppong steadily. The big old 300-watt supplies that
were designed for running several full-height 5.25" floppies are overkill in
these days of portable-driven lightweight 3.5" drives. 200 watts is good
enough these days, and the new breed of compact 200W supplies is quieter to
boot.
(About that annoying fan noise, ask if the fan on a target system has a
variable speed motor with thermostatic control --- this will cut down on noise
tremendously. If not: I have seen a rave about, but haven't used, a
thermostatic fan controller called "The Silencer". This tiny device mounts
inside your power supply and connects to the fan's power leads. It
automatically varies the fan motor speed to hold a 79 to 82F temperature.
Write Quiet Technology, Inc. PO Box 8478, Port St. Lucie FL 34985. Warning:
installing this may void your warranty!)
C. The Heart Of The Machine
Yet another basic decision, of course, is processor speed and type. Until
recently, the hot sellers in this market were the 386/33DX and AMD 386/40DX,
which I'd say are reasonable minimum-speed engines for UNIX with X. However,
recent pricing moves by Intel have moved the price of a 486SX25 below the
equivalent 386DX33 chip. The 386 is therefore effectively dead for new
hardware, and the 486SX/25 defined as the new low end (at least for the next
90 days or so).
The 386SX machines were never a very good idea for UNIX; the 16-bit bus-to-CPU
path can choke your throughput. The 486SX is even worse, a stupid marketing
crock with no technical justification whatsoever. It's a 486DX with the
floating-point unit missing or even deliberately lobotomized out; the
difference *will* bite you in unobvious ways, for example if you use X which
does a lot of FP for graphics. (One respondent opined that "SX" is
Intel-internal code for "sucks".)
A 486DX/33 has enough power make a good personal UNIX box. For UNIX, this is
your floor; how far above them you want to buy depends on your budget and job
mix. We'll have much more to say about this in the section on performance
tuning.
D. Motherboards and BIOSes
Provided you exercise a little prudence and stay out of the price basement,
motherboards and BIOS chips don't vary much in quality either. There are only
six or so major brands of motherboard inside all those cases and they're pretty
much interchangable; brand premiums are low to nonexistent and cost is strictly
tied to maximum speed and bus type. Unless you're buying from a "name" outfit
like Compaq, Dell, or AST that rolls its own motherboards and BIOSes, there are
only four major brands of BIOS chip (AMI, Phoenix, Mylex, Award) and not much
to choose between 'em but the look of the self-test screens. One advantage
UNIX buyers have is that UNIXes are built not to rely on the BIOS code (because
it can't be used in protected mode without more pain than than it's worth). If
your BIOS will boot properly, you're usually going to be OK.
If the above sounds too rosy, there is a catch; it describes *current*
hardware, not some of the historical botches. And it's hard to know how old
what you're buying is. You might actually be buying a motherboard that's been
sitting on the dealer's back shelf for a year, with a BIOS chip in it that was
in the drawer for another year before he ever stuck it in the board. And some
of those older BIOSes and board designs are to be desperately avoided. There
have been quite a few bogus cache designs that either don't work at all
(instant panic under UNIX) or that severely degrade performance. A lot of
earlier designs have bus timing problems that show up in bad interactions with
host adapters and fancy graphics boards. Bad memory designs were also not
uncommon.
A good, tricky way to keep the vendor from shipping you these fossils is to
specify a motherboard that can take 4MB SIMMs (as opposed to just the older 1MB
kind). You want to do this anyhow for functional reasons.
There are a few other potential gotchas to beware of, especially in the cheaper
off-brand boards. One is "shadow RAM", a trick some boards use for speeding up
DOS by copying the ROM contents into RAM at startup. It should be possible to
disable this. Also, on a caching motherboard, you need to be able to disable
caching in the memory areas used by expansion cards. Some cheap motherboards
fail to pass bus-mastering tests and so are useless for use with a good SCSI
interface; on others, the bus gets flaky when its turbo (high-speed) mode is
on. Fortunately, these problems aren't common.
Finally, one name-brand tip: *don't* buy DTK-brand motherboards for a UNIX
system! They generate lots of spurious interrupts, which DOS is too stupid to
be bothered by but which completely tank UNIX.
You can avoid both dangerously fossilized hardware and these little gotchas by
sticking with a system or motherboard design that's been tested with UNIX (some
help with that below).
Some other good features to look for in a motherboard include:
* Gold-plated contacts in the expansion slots and RAM sockets. Base-metal
contacts tend to grow an oxidation layer which can cause intermittent
connection faults that look like bad RAM chips or boards. (This is why, if
your hardware starts flaking out, one of the first things to do is jiggle
or remove the boards and reseat them, and press down on the RAM chips
to reseat them as well -- this may break up the oxidation layer. If
this doesn't work, rubbing what contacts you can reach with a soft
eraser is a good fast way to remove the oxidation film. Beware, some
hard erasers, including many pencil erasers, can strip off the plating, too!)
* Ability to go to 64MB on the motherboard (that is, without plug-in
daughterboards). Most EISA boards seem to have this (the popular Mylex
MAE486 board is an exception).
* The board should be speed-rated as high as your processor, of course.
It's good if it's rated higher, so upgrade to a faster processor is
just a matter of dropping in the chip and a new crystal.
If your motherboard offers multiple cache sizes, make sure you know whether the
larger cache is required when using more than a certain amount of memory. Or,
in general, fill the cache all the way -- cache-speed RAM is getting pretty
cheap.
Note, however, that hardware caches for system boards are really designed to
achieve effective 0 wait state status, rather than perform any significant
buffering of data. As a general rule applicable to all clones, 64Kb cache
handles up to 16Mb memory. 256Kb cache handles up to 64Mb. 128Kb cache is
nearly redundant for a 16MB system; the benefit from additional caching is
statistically. This means that running with 8Mb RAM, there is little
difference between a 64Kb, 128Kb, or 256Kb cache on the systemboard.
Finally, beware the infamous FP exception bug! Some motherboards fail to
handle floating point exceptions correctly; instead of generating a SIGFPE they
lock up. The following fragment of C code will reproduce the problem:
double d;
d = 0.0;
d = 1.0 / d; /* floating divide by zero should yield SIGFPE */
John R. Levine <jo...@iecc.cambridge.ma.us> explains: "The difficulty stems
from the fact that there are two ways to handle floating exceptions on a 486,
the right way and the PC way. What the 486 wants to do is to generate an
interupt 16 when there is a floating point error, all entirely internal to the
CPU. This has been the native way to handle floating point interrupts since
the 286/287. The 286/287 and 386/387 each have a dedicated ERROR pin that the
FPU uses to tell the CPU that it's time for an error interrupt.
Unfortunately, the 8086/8087 handled interrupts differently. The error pin on
the 8087 was wired to the 8259A interrupt controller, the same interrupt
controller that handled keyboard, disk, clock, etc. interrupts. The PC/AT
enshrined IRQ 13 as the one for floating interrupts. (The details of this are
a little hazy to me, since the XT didn't have IRQ 13 tied to an 8259A, so the
AT must have at least changed the interrupt number.) PC designs have generally
wired the 287 or 387 ERROR pin to the 8259A, not to the ERROR pin on the CPU,
or at best had some poorly documented way to switch between the two interrupt
methods.
In the interest of backward compatibility, the 486 has a mode bit that says not
to handle FP exceptions automatically, but rather to freeze the FPU and send a
signal on the FERR pin, which is usually tied to an 8259A which then feeds the
interrupt back as IRQ 13. There is some extra complication involved here
because the FPU has to stay frozen until the interrupt is accepted so the CPU
can go back and look at the FPU's state. Early 386/25 chips had a bug that
would sometimes freeze up on a floating point interrupt and you had to get a
kludge socket with a PAL that fixed the timing glitch that provoked the bug.
So as likely as not, the motherboard hardware that runs FERR out and back isn't
working correctly. It's not surprising, few DOS users take floating point
seriously enough to notice whether the interrupts are working right."
When you specify a system, make clear to your vendor that the motherboard must
handle float exceptions properly. Test your motherboard's handling of
divide-by-zero; if it doesn't work, press your vendor to replace it *and
send me email*! Only by publishing a list of boards known bad can we
protect ourselves and pressure vendors to fix this problem.
The 386 UNIX Buyer's Guide posting (pc-unix/software) includes tables of
motherboards and systems known to run with various UNIX ports.
E. Peripherals
Peripherals are another matter, especially hard disks. A good rule of thumb
for balanced configurations is that the hard disk should comprise about half
(or maybe a bit more) of the total system hardware price (exception: if you're
buying a really good monitor, like 16" or over, it's going to be expensive
enough to bust this rule). Unless you're the exception who has to invoke
warranty due to a system arriving dead, most of what you buy from a dealer or
mail-order house is their ability to surf the Winchester market, make volume
buys, and burn in your disks before shipping. We'll look at disk choices in
more detail later on.
These days, most vendors bundle a 14" monitor and super-VGA card with 1024x768
resolution in with their systems. Details to watch are whether the card comes
loaded with 512K or 1MB of RAM (which will affect how much of that maximum
resolution and how many colors you actually get), whether the memory is
dual-ported VRAM (slightly more expensive but much faster), and whether the
monitor is interlaced or non-interlaced. The latter is better and should no
longer cost extra; look for the abbreviation NI in the ad or quote and be
suspicious if you don't see it.
You should have a tape drive for backup. Unfortunately, the tape drive market
is rather confusing. Rather than try to give a capsule summary, we give it
its own section below.
We'll have much more to say about price/performance tradeoffs in peripherals
in the next major section, on performance tuning.
F. Keyboards
Hal Snyder of Coherent, Inc. <h...@mwc.com> sent us the following caveat:
We find that about 10% of cheap no-name keyboards do not work in scan
code set 3. We are interested in scan code set 3 because only there can
you reprogram the keyboard on a per-key basis as to whether keys are
make-only, make-break, or autorepeat. It is a big win for international
support and for X.
Keytronic, Cherry, and Honeywell keyboards, as well as a large number of
imports, work fine. My advice is to either by a respected brand of
keyboard, or deal with a vendor who will allow you to return an
inompatible keyboard without charge.
G. Power Protection
Finally, I strongly recommend that you buy a power conditioner to protect your
hardware. MOV-filtered power bars make nice fuses (they're cheap to replace),
but they're not enough. I've been delighted with my TrippLite 1200, which you
can get for $139 or so by mail order. A fringe benefit of this little beauty
is that if you accidently pull your plug out of the wall you may find you
actually have time to re-connect it before the machine notices!
The tecbical info in the remainder of this section is edited from material
supplied by David E. Wexelblat <d...@mtgzfs3.att.com>.
There are several levels of power protection available to the home computer
user. I break this down into 4 levels; others may have different ways of
classifying things. The levels are:
1. Surge Suppressor
2. Line Conditioners
3. Standby Power Supplies
4. Uninterruptible Power Supplies
and here's what they mean:
1. Surge suppressors
These are basically a fancy fuse between the source and your hardware; they
clamp down spikes, but can't fill in a low voltage level or dropout.
This is a bare minimum level of protection that any piece of expensive
electronics should have. Note that this applies to more than just AC power;
surge suppressors are available for (and should be used on) phone lines, and
RS-232 and parallel connections (for use on long lines; generally not needed if
the devices is colocated with the computer and all devices are protected from
outside sources). Note also that *all* devices connected to your computer need
to be protected; if you put a surge suppressor on your computer but not your
printer, then a zap on the printer may take out the computer, too.
An important fact about surge suppressors is that *they need to be replaced if
they absorb a large surge*. Besides fuses, most suppressors rely on on
components called Metal-Oxide Varistors (or MOVs) for spike suppression, which
degrade when they take a voltage hit. The problem with cheap suppressors is
that they don't tell you when the MOV is cooked, so you can end up with no
spike protection and a false sense of security --- better ones have an
indicator.
You can buy surge suppressors at any Radio Shack; for better prices, go
mail-order through Computer Shopper or some similar magazine. All of
these are low-cost devices ($10-50).
2. Line Conditioners
These devices filter noise out of AC lines. Noise can degrade your power
supply and cause it to fail prematurely. They also protect against short
voltage dropouts and include surge suppression.
My Tripp-Lite 1200 is typical of the better class of line conditioners --- a
box with a good big soft-iron transformer and a couple of moby capacitors in it
and *no* conductive path between the in and out sides. With one of these, you
can laugh at brownouts and electrical storms.
Price vary widely, from $40-400, depending on the power rating and capabilities
of the device. Mail-order from a reputable supply house is your best bet.
Line conditioners typically *don't* need to be replaced after a surge; check
to see if yours includes MOVs.
3. Standby power supplies (SPSs)
These devices are battery-based emergency power supplies that provide power for
your system via an inverter if the power fails. An SPS will generally have all
the capabilities of a line conditioner as well.
Note: these devices do not come on line until after the power fails, and have a
certain amount of delay of some of milliseconds before they come on line. If
the capacitors in your power supply are not large enough, the SPS may not cut
in in time to prevent your computer from seeing the power failure.
Note also that many SPSs are marketed as Uninterruptable Power Supplies (see
below). This is incorrect. Any device with a non-zero cutover time cannot be
a true UPS. If the ad mentions a cutover time, it's an SPS, and not a UPS.
The price range for these devices (depending largely on size and cutover time)
is $200-2000. An SPS will *not* need to be replaced after absorbing a large
surge.
4. Uninterruptable power supplies (UPSs)
These devices provide full-time isolation from the incoming AC line through a
transformer of some sort. These devices are on-line at all times, and if the
AC line fails, the batteries will cut in. Your devices will see no
interruption of their incoming AC. UPSs cost more, and provide more features.
They are the ultimate in power protection. Many UPSs have an intelligent
interface that will notify a connected device of a power failure, allowing it
to shut down cleanly. UPSs also provide the capabilities of a line
conditioner. The price range (for devices in the size range for a home
computer) are $400-$2500. An UPS will *not* need to be replaced after
absorbing a large surge.
Now, given this information, how does one decide what to get? For a system
that runs unattended, like most Unix systems, it is best to have a device that
provides both power holdover and a power failure signal. Hence, for a Unix
system, a UPS is the best idea (an SPS is not the best power protection and
most have no intelligent interface). At least one vendor sells
ordered-shutdown software for Unix, and it's fairly simple to write your own
daemon to monitor a serial port, and send init a SIGPWR signal when it sees a
certain signal.
Our recommendation for a home Unix environment is a configuration like the
following:
a) A true on-line UPS for the computer system. An intelligent
interface is mandatory, along with appropriate software for
ordered shutdown.
b) Surge suppression on all phone lines, and also on serial/parallel
lines that leave the room.
c) Line conditioners on any devices not connected to the UPS. If
you do take a power hit, it's cheaper to replace a $50 line
conditioner than a $1500 laser printer.
An important question is "How do I know how big a UPS to get?" The watt rating
of the UPS should be at least the sum of the peak ratings off all equipment
connected to it. UPS marketroids tend to quote you UPS capacities and formulas
like "sum of VA ratings + 20%" which (surprise!) push you towards costler
hardware. Ignore them. If a watt rating is not given, watts = 0.75*VAmax.
One other consideration is that you typically shouldn't put a laser printer on
a UPS --- toner heaters draw enough current to overload a UPS and cause a
shutdown within seconds. The other thing is that you can't even put the laser
printer on the same circuit with a UPS --- the heater kicks on every 20-30
seconds, and most UPSs will see the current draw as a brownout. So buy a
separate line conditioner for the laser printer.
Finally, read the UPS's installation manual carefully if you're going
to use it with other power-protection devices. Some UPSs don't like having
surge suppressors between them and the equipment.
David personally recommends surge suppressors and line conditioners from
Tripp-Lite (available both mail-order and retail), and UPSs from Best Power
Technologies (Necedah, WI - 1-(800)-356-5737). I can enthusiastically second
the TrippLite recommendation, but haven't dealt with Best Power at all. There
are many other vendors for all of these devices.
Tripp-Lite has a whole range of products, from a $10 phone-line
surge-suppressor, to line conditioners and SPSs with prces in the hundreds of
dollars. They have a line of $50-80 line conditioners that are good for most
peripherals (including your home stereo :->).
Best Power Technologies sells two lines of UPSs in the range for home systems.
The older and more expensive FERRUPS line (which is what David has) has a smart
interface, and very good filtering and surge-suppression capabilities. He says
"I have a 1.15kVA FERRUPS for my home system, which is overkill with my current
hardware (although it rode out a 45 minute power failure with nary a whisper -
no reboot). In 1990, I paid ~$1600 for this device, and that has since gone
up. They also sell a newer line of Fortress UPSs. These are better suited in
price for home systems. I don't know much about them, as they were not
available when I bought my UPS. I expect that this is what most people will
want to consider, though. In addition, Best sells Check-UPS, a software
package (in source form) for monitoring the UPS and shutting it down. I have
found Best to be a good company to deal with, with competent, knowledgable
sales people (who will be able to help you pick the right device), and helpful,
courteous, and responsive technical support."
Other things to know:
A UPS should be wired directly to (or plugged directly into) the AC supply
(i.e. a surge suppressor is neither required nor suggested between the wall and
the UPS). In addition, a surge suppressor between the UPS and the equipment
connected to it is redundant and also unnecessary.
IV. Performance tuning
Here are the places where you can trade off spending against the performance
level you want to buy and your expected job mix.
A. How to Pick your Processor
The following information appeared in article <13a29iINN...@iraul1.ira.uka.de>
by S_JU...@iravcl.ira.uka.de (|S| Norbert Juffa). It gives a good indication
of the relative speeds in Intel's processor line:
UNIX performance of Intel processors as given in Intel's literature
Processor SPECmark SPECint SPECfp Whetstone Dhrystone Linpack Ref Rm
double p. 2.1 dp MFLOPS
1) Intel 386/387-33 4.3 6.4 3.3 3290 15888 N/A 1 *+
2) Intel 386/387-33 4.1 6.0 N/A 3200 18900 0.4 2 #
3) RapidCAD-33 6.6 7.3 6.1 5300 18275 N/A 1 *+
4) 486DX-25 8.7 13.3 6.6 5640 32000 1.0 2
5) 486DX-33 11.1 17.5 8.2 7200 43000 1.5 3
6) 486DX-33 12.1 18.3 9.2 N/A N/A N/A 4
7) 486DX-33 14.5 19.0 12.2 12300 43500 1.6 5 &
8) 486DX-50 18.2 27.9 13.6 10710 64400 2.5 3
9) 486DX2-50 19.2 25.4 15.9 18500 63966 2.3 5 &
10)486DX-50 21.9 28.5 18.3 18500 65400 2.4 5 &
11)486DX2-66 25.6 34.0 21.2 24700 85470 3.1 5 &
Remarks:
* Whetstone/Dhrystone are 32-bit DOS results
+ SPEC ratios recomputed from SPEC timings (computed wrong in report)
& note huge increase in SPEC floating point performance over previous results
due to new experimental FORTRAN compiler
# machine with AMD 386-40/Cyrix 83D87-40/128 kB cache is estimated by me at:
7.7 SPECint, 5.0 SPECfp, 6.1 SPECmark,
5600 double prec. Whetstones, 23000 Dhrystones,
0.6 Linpack double prec. MFlops
These estimates based on my own measurements and data from:
FasMath 83D87 Benchmark Report, Cyrix 1990
World's Fastest 386 40 MHz Am386(tm)DX Microprocessor Performance Summary,
AMD 1991
References:
1) Intel RapidCAD(tm) Engineering CoProcessor Performance Brief. 1992
2) i486(tm) Microprocessor Performance Report. 1990.
Order No. 240734-001
3) 50MHz Intel486(tm) DX Microprocessor Performance Brief. 1991.
Order No. 241120-001
4) i486(tm) Microprocessor Business Performance Brief. 1990.
Order No. 281352-002
5) Intel486(tm) DX2 Microprocessor Performance Brief. 1992
Order No. 241254-001
Configurations:
1) COMPAQ SystemPro 386/33 MHz, 8 MB memory, AT&T UNIX System V/386 Release 4.0
Version 2.0
2) 64 kB write back cache,
AT&T UNIX System V Release 3.2CC, MetaWare High C R2.2c,
SVS FORTRAN V2.8
3) COMPAQ SystemPro 386/33 MHz, 8 MB memory, AT&T UNIX System V/386 Release 4.0
Version 2.0
4) 128 kB write-back cache, 12 MB RAM,
AT&T UNIX System V Release 3.2CC, MetaWare High C R2.2c,
SVS FORTRAN V2.8
5) No 2nd level cache, 16 MB RAM,
AT&T UNIX System V/386 R3.2, MetaWare High C R2.3p
SVS FORTRAN V2.8
6) ALR PowerCache 33/4e, 128 kB cache, 16 MB RAM
SCO UNIX System V R3.2.2, MetaWare High C R2.2c/R2.3k,
SVS FORTRAN V 2.8
7) Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
8) 256 kB write-back cache (82495DX/82490DX), 16 MB RAM,
AT&T UNIX System V/386 R3.2, MetaWare High C R2.3p
SVS FORTRAN V2.8
9) Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
10)Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
11)Intel Modular Platform, 256 kB write-back cache, 32 MB RAM,
AT&T UNIX System V R4.0.4, Metaware High C R2.4b,
Intel Scheduling FORTRAN 77 Compiler V0.2
One of Intel's most recent wrinkles is the "clock-doubler" chips. The 50DX2
runs at 25MHz externally but computes at 50MHz. A 66DX2 (bus speed 33MHz) is
also shipping, and there are persistent rumors of a clock-doubled 50 in the
works that would compute at a blistering 100MHz! Intel likes to claim a 70%
speedup for the doublers over their undoubled brethren. I've expressed
skepticism about this in previous issues, but the SPECmarks above suggest that
just this once the marketroids may not be lying -- much. Under UNIX, a 50DX2
is in fact nearly as fast as a true 50DX. Still, beware of anyone whose
literature passes off the DX2 qualification in the fine print; they may be
scamming about other things, too.
Right now you'll pay as much as a $500 premium for a 486/50, as that's
relatively new technology and demands extra-fast memory to run full-out. Also,
these processors run really hot (one correspondent described the 50 as a
"toaster on a chip"). If you go this route, be sure your configuration has an
extra-heavy-duty cooling fan. Or two. And, for preference, a hefty heat sink.
Of course, if you do this you'll be ready to drop in the rumored 100DX2 part,
and blow the doors off all those fancy proprietary-technology workstations.
B. Of Memory In...
Buy lots of RAM, it's the cheapest way to improve real performance on any
virtual-memory system. At $30-$50 maximum per megabyte it's just plain silly
to stick with the 2-4mb now standard on most clone configurations. Go to 8,
you won't regret it; 16 if you're going to use X.
Above 16 is iffy on ISA boxes because the stock USL 4.0.3 kernel may try to do
DMA from a location the bus can't deal with. Most UNIX vendors have fixed this
by adding code that forces DMAs to take place from low memory; make absolutely
sure that includes yours before you load up beyond 16MB. The pc-unix/software
FAQ posting includes information on which vendors are known to have fixed this
problem.
Some motherboards have 16 sockets for SIMM memory modules. Some only 8. Some
take only 1MB mdules, some handle 4MB. These constraints interact in funny
ways.
You should make sure if you are buying an entry level 2 or 6 MB system with a
16-socket motherboard that you will not have to ditch the SIMMs that are
already installed in order to go to your maximum (if 16 MB is your maximum).
Some systems only allow you to mix 1M and 4M SIMMs in certain combinations.
Try not to get any 1M SIMMs in your initial configuration, because you'll
probably end up turfing them later. That is, buy a 4MB, 8MB, 12 MB or 16MB
system to start.
Newer ISA designs have a 32 MB upper limit with only 8 sockets, since they can
take 4Mx9s...however, this means different interleaving (only 2 banks), which
limits the possible configurations. You don't want to start off with an 8 MB
configuration, because that's 8 ea 1Mx9's, filling up all the sockets...the
next upgrade requires replacing 1Mx9 with 4Mx9. You can't even set up 12
MB!...the first reasonable config (that won't require tossing hardware) is 16
MB, since that's one bank full of 4Mx9.
Most EISA motherboards have 16 4MB-capable sockets, and this is clearly
where the market is going.
C. Bus wars
Should you buy 16-bit ISA vs. 32-bit EISA? You'll pay a $600-$900 premium for
the latter. What you get in return is the ability to use things like fast
32-bit SCSI controllers and a smoother upward-migration path. On the other
hand, EISA cards are significantly more expensive. And so far, there isn't
much support for EISA-specific hardware --- a couple of vendors will drive EISA
SCSI disk and tape controllers and that's about it (of course those *are* the
most important bandwidth-eaters). All ISA cards will still work.
Of course, most of what you get from EISA is a performance boost. There are
two different theories about why EISA is better; both have their adherents.
Theory A: Bandwidth matters
UNIX has always been an I/O-intensive operating system. According to this
theory, increasing processor speed on clones can leave it spending all its time
waiting on the limited I/O capacity of the poor old 5.3MB/sec ISA bus. The
vendors all seem to think this starts at around 33MHz and that if you're buying
50MHz it definitely pays to go EISA.
Theory B: Cache is what matters
According to this theory, UNIX never comes even close to saturating the ISA-bus
bandwidth. EISA boards are faster because the premium vendors can charge for
them allows the motherboard designer more freedom and a richer parts budget.
The most important performance effect of this is that EISA boards have larger
and better-designed caches, increasing the effective memory-access speed.
There's probably some truth to both analyses. If your machine is going to
spend most of its processor time running X displays and doing other classically
compute-bound tasks, cache size matters most. On the other hand, benchmarks
show that the combination of TCP/IP and multi-user disk activity *can* saturate
ISA, and one can sometimes *see* a fast-processor machine slow down during disk
accesses...
If you're contemplating any kind of heavy-duty networking, EISA network
adapters will become rather important. A correspondent tells me he's seen
benchmarks showing what percentage of bus bandwidth is consumed by various
cards when flooding an ethernet (i.e. consuming the entire 10Mbit bandwidth of
a quiet net, as you might be when doing an FTP transfer, for instance). 8-bit
ISA cards consume 40-60% of bus bandwidth; 16-bit cards, 20-40%. 32-bit EISA
cards consume only about 5-10%. This would be particularly important in a
machine being used as a bridge, where you might be handling a large portion of
the traffic on two or more separate nets. The advantage of EISA cards may be
due to their shorter-cycle bus mastering DMA. At time of writing, only
SCO supports these, but other UNIX vendors are known to have their own drivers
in the pipeline.
D. IDE vs. SCSI
Another basic decision is IDE vs. SCSI. Either kind of disk costs about the
same, but the premium for a SCSI card varies all over the lot, partly because
of price differences between ISA and EISA SCSI cards and especially because
many motherboard vendors bundle an IDE chip right on the system board. SCSI
gives you better speed and throughput and loads the processor less, a win for
larger disks and an especially significant consideration in a multi-user
environment; also it's more expandable.
Another important win for SCSI is that it handles multiple devices much more
efficiently. If you have two IDE (or ST506 or ESDI) drives, only one can
transfer between memory and disk at once. In fact, you have to program them at
such a low level, one drive might actually be blocked from *seeking* while
you're talking to the other drive. SCSI drives are mostly autonomous and can
do everything at once; and current SCSI drives are not quite fast enough to
flood more than 1/2 the SCSI bus bandwidth, so you can have at least two drives
on a single bus pumping full speed without using it up. In reality, you don't
keep drives running full speed all the time, so you should be able to have 3-4
drives on a bus before you really start feeling bandwidth crunch.
All this having been said, don't write off IDE too quickly. Sure, it's
compatible with the nasty old ST506 interface, but it's *much* faster. It
remains the cost-effective choice for smaller drives (up to 500MB) on systems
that won't be hitting the disk constantly. Unless you're running a heavily
used network or database server, don't assume SCSI will make any noticeable
difference.
(If you can do your own installation, I hear that used 150/250MB SCSI drives
are getting quite common and cheap on the net. All 150MB QIC type drives can
do 250MB on extended-length tapes, though some manufacturers discourage you
from doing this to avoid excessive heade wear. But back to disks...)
The following, by Ashok Singhal <ash...@duckjibe.eng.sun.com> of Sun
Microsystems, is a valiant attempt to demystify SCSI terminology.
The terms "SCSI" and "SCSI-2" refer to two different specifications.
Each specification has a number of options. Many of these options are
*independent* of each other. I like to think of the main options (there are
others that I'll skip over because I don't know enough about them to talk
about them on the net) by classifying them into five categories:
1. Logical
This refers to the commands that the controllers understand.
SCSI-2 defined a common cammand set that is pretty much a
superset of the SCSI command set.
2. Data Width
8 bits (+ 1 parity) -> "normal"
16-bits (+ 2 parity) -> "wide"
32-bits (+ 4 parity) -> I don't know, "extra-wide??"
All three options are available in SCSI-2 (yes,
the draft spec I have even shows 32-bits!), although
8-bit wide is still by far the most common. Not sure, but I believe
SCSI defined only 8-bit wide data path.
3. Electrical Interface
single-ended (max cable length 6 meters)
differential (max cable length 25 meters)
Both options are available for SCSI-2 (I'm not sure about SCSI,
but I think both options were available also)
and this option is independent of options 2, 4, 5. Differential
is less common but allows better noise immunity and longer
cables.
4. Handshake
Asynchronous (requests and acks alternate)
Synchronous (multiple requests can be outstanding)
Both options are available for SCSI-2 (Not sure about SCSI,
but I think both were available also). This is negotiated
between each target and initiator; asynchronous and synchronous
transfers can occur on the same bus. This is independent of
2, 3 (Not sure about 1).
5. Synchronous Speed (does not apply for asynchronous option)
"Normal" is up to 5 Mtransfers/sec ( = 5MB/s for 8-bit wide, more
for wider)
"Fast" is up to 10 Mtransfers/s ( = 10 MB/s for 8-bit wide, more
for wider)
The fast option is defined only in SCSI-2.
This options basically defines shorter timing parameters
such as the assertion period and hold time.
The parameters of the synchronous transfer are negotiated
between each target and initiator so different speed transfers
can occur over the same bus.
E. Other Disk Decisions
Look at seek times and transfer rates for your disk; under UNIX disk speed and
throughput are so important that a 1-millisecond difference in average seek
time can be noticeable.
Previous issues said "Disk caching is good, but there can be too much of a
good thing. Excessively large caches will slow the system because the overhead
for cache fills swamps the real accesses (this is especially a trap for
databases and other applications that do non-sequential I/O). More than 100K
of cache is probably a bad idea for a general-purpose UNIX box; watch out for
manufacturers who inflate cache size because memory is cheap and they think
customers will be impressed by big numbers." This may no longer be true on
current hardware; in particular, most controllers will interrupt a cache-fill
to fulfill a `real' read request.
In any case, having a large cached hard drive (particularly in the IDEs) often
does not translate to better performance. For example, Quantum makes a 210Mb
IDE drive which comes with 256Kb cache. Conner and Maxtor also have 210Mb
drives, but only with 64Kb caches. The transfer rate on the drives, however,
show that the Quantum comes in at 890Kb/sec, while the Maxtor and Conner fly
away at 1200Kb/sec. Clearly, the Conner and Maxtor make much better use of
their smaller cache.
Many retailers seem to enjoy advertising the "9ms" Quantum 52/80/120/200Mb
drives. This speed, of course, is bogus. All the quantum drives are at least
16ms is average access. The 9ms already includes the cacheing speedup.
However, it may be that *any* hardware disk caching is a lose for UNIX! Scott
Bennett <benn...@mp.cs.niu.edu> reports a discussion on comp.unix.wizards:
"nobody found the hardware disk caches to be as effective in terms of
performance as the file system buffer cache...In many cases, disabling the
hardware cache improved system performance substantially. The interpretation
of these results was that the caching algorithm in the kernel was superior to,
or at least better tuned to UNIX accesses than, the hardware caching
algorithms."
Thus, if your disk controller allows it, try disabling the cache. Your
throughput may go up!
F. Souping Up X Performance
One good way to boost your X performance is to invest in a graphics card with a
dedicated blitter or high-speed local-bus connection, like the ATI series or
the S3-based Quantum, Wind/X and Orchid Fahrenheit 1280. A number of clone
vendors offer these accelerator options relatively cheap and can make your X go
like a banshee; however, stock X doesn't support them yet. Third-party servers
are available from MetroLink (email sa...@metrolink.com) or SGCS
(i...@sgcs.com).
Here is a current price list from MetroLink:
Runtime (all servers, standard and contrib clients) : 299.00
Development (full X11 and Motif 1.1.4 libraries) : 299.00
Xv - Real-Time Video in an X window (true server : 99.00
extension)
Xie - X Imaging Extension : 199.00
And here is the corresponding info from SGCS:
Ref # Description Price
----- --------------------------------------------- ------
** 1 Full X11R5 binaries licensed for a single CPU 295.00
2 Additional X11R5 binary license 50.00
** 3 Enhanced X11R5 source code (note 4) 195.00
4 MIT source code of contributed clients (note 5) 50.00
** 5 Motif binaries for a single CPU 245.00
6 Additional Motif binary license 60.00
** 7 X11R5 Documentation Set 150.00
8 PHIGS Documentation Set 75.00
** DISCOUNTS:
If your choose more than one selection from any of the (**) items above
you will receive the following discounts: $50 off on 2 selections,
$75 off on 3 selections, $100 off on 4 selections
In general, the ATI approach (normal bus, dedicated blitter and optimization
for special functions like character drawing) will speed up text display, text
scrolling and window resize/move operations a lot, but line-drawing and
graphics only a little. S3, on the other hand, speeds up high-bandwidth
graphics drawing a lot but doesn't have as big an advantage for ordinary
text operations. You pays your money and takes your choice. Benchmarks
indicate that most non-CAD users are better served by the ATI approach.
If you're feeling *really* flush, plump for a 15", 17" or even 20" monitor.
The larger size can make a major difference in viewing comfort. Also you'll be
set for VESA 1280x1024 when everybody gets to supporting that. In the mean
time, the bigger screen will allow you to use fonts in smaller pixel sizes so
that your text windows can be larger, giving you a substantial part of the
benefit you'd get from higher pixel resolutions.
If you can, buy your monitor from someplace that will let you see the same
monitor (exact same, not the same monitor) that will be on your system.
There's a *lot* of quality variation even in "premium" monitor brands.
The VESA (Video Electronics Standards Association) standard for local bus video
connectors is now out. When you buy local-bus motherboards, insist that they
be VESA-conforming. Be very clear about this and get a commitment from your
vendor; some unscrupulous operations may still be attempting to unload pre-VESA
motherboards on unsuspecting customers.
V. Tape Drive Follies
You should have a tape drive for backup, and because most UNIX vendors like to
distribute their OS on tape. Ideally, your tape backup should be able to image
your entire disk. Unfortunately, this can get quite expensive for large disks,
as we'll see below.
There are two major technologies in today's desktop tape drive market; QIC
(Quarter Inch Cartridge) at the low end and midrange, and DAT (Digital Audio
Tape) at the high end. The dividing line is about 1MB capacity.
DAT is a new technology; it's not far down its price curve yet, but clearly
where the future is. DAT drive capacities are quoted in *gigabytes* (that is,
thousands of megabytes).
Conventional QIC drives have capacities up to 525 megabytes (a little more than
half a gig). They are a mature technology, but one plagued by hardware
incompatibilities and driver bugs. Part of the problem is that, until
recently, hard disks were small enough relative to a floppy's capacity that
demand for high-volume backup technology was low in the PC world; QIC vendors
tended to be small, insular, technology-driven firms relatively uninterested in
standardization.
As a result, understanding tape drive specifications is far from trivial.
Tape drive standards are develeped by Quarter Inch Cartridge Drive Standards,
Inc. (805-963-3853), a consortium of drive and media vendors. They develop
standards for controllers, transports, heads, and media. Some of these
become ANSI standards. We'll discuss the most iomportant ones here.
Common Tape Drive Interfaces:
QIC-02 --- intelligent hardware tape interface
QIC-36 --- simple hardware tape interface
QIC-104/11 --- SCSI-1 tape interface
QIC-121 --- SCSI-2 tape interface
These standards describe the drive controller. QIC-02 is presently by far the
most common, and QIC-36 nearly obsolete (it was designed at a time when
on-board intelligence for controllers was much more expensive than now). The
SCSI standards are only rarely cited by number; usually, QIC-104 and QIC-121
devices are referred to simply as "SCSI tapes".
Common Recording formats:
QIC-24 --- 9-track 60-Mbyte tape format
QIC-120 --- 15-track 125-Mbyte tape format
QIC-150 --- 18-track 150-Mbyte tape format
QIC-525 --- 26-track 525-Mbyte tape format
These standards describe the drive itself.
Now, in theory, these standards are upward compatible; that is, a QIC-120 drive
can read a QIC-24 tape, a QIC-150 drive can read both QIC-120s and QIC-24s, and
so on. There's a potential gotcha here, though, called "media
incompatibility". Thus, we also need to consider:
Common media:
DC600A --- for QIC-24 and QIC-120 drives
DC6150 --- for QIC-150 drives
DC6525 --- for QIC-525 drives
The Wangtek 5150ES (and possibly some other 525-megabyte drives) will,
according to its documentation, decode QIC-24 --- but it won't read a DC600A
medium!
So, make sure your tape drive can read the media your OS vendor is going to
ship on. QIC-24 on DC600As and QIC-150 on DC6150s are very widely used as a
software distribution format in the UNIX world, and you probably want to make
sure your drive can read them.
60/120MB QIC drives are fairly cheap now but larger sizes (typically 150, 250,
525 QIC tapes and 1.3gig DAT) are not. DAT drives, in particular, cost more
than a grand each (however, if you have large drives the cost difference can
quickly get weaten up by media costs).
One interesting point is that if you've gone SCSI, a 150MB QIC (comparable to
the drives now popular on Suns) may well be cheaper than older 60MB technology;
the win is in the controller prices, which have plummeted since QIC-24 was the
cutting edge.
Tape drives are easy to find and pretty safe to buy through mail order. It's
also possible to buy reconditioned but warrantied used drives substantially
cheaper than new. One correspondent recommended Super Technologies of Chino,
CA (800 322 3999); they'll sell you a rebuilt Wangtek 150 with a 7-month
warranty and a controller card for $300 and change, or a DAT drive for $800.
Ron Swanson 612-733-9458
Dale Huff 612-736-0551
VI. Of Mice and Machines
In a previous issue, I claimed that all mice and trackballs are the same for
compatibility purposes. I was wrong -- seriously wrong. The more I found out,
the messier the picture gets. The following is an attempt to sort out all the
confusion. Thanks to Jim McCarthy at Logitech for digging into the matter
and somewhat alleviating my ignorance.
Mice and trackballs used to be simple; now, thanks to Microsoft, they're
complicated. In the beginning, there was only the Mouse Systems 3-button
serial mouse; this reported status to a serial port 30 times a second using a
5-byte serial packet encoding now called "C" protocol. The Logitech Series 7
and 9 mice were Mouse Systems-compatible. All UNIXes that have any mouse
support at all understand C-protocol serial mice.
Then Microsoft got into the act. They designed a two-button serial mouse which
reports only deltas in a three-byte packet; that is, it sends changes in button
status and motion reports only when the mouse is actually moving. This is
called `M' protocol. Microsoft sold a lot of mice, so Logitech switched from
`C' to `M' --- but they added a third button, state changes for which show up
in an optional fourth byte. Thus, `M+' protocol, upward-compatible with
Microsoft's `M'. Most UNIX vendors add support for M+ mice, but it's wise to
check.
Bus mice are divided into 8255 and InPort types. These report info
continuously at 30 or 60 Hz (though InPort mice have an option for reporting
deltas only), and you get interrupts on events and then have to poll hardware
ports for details. More on these next issue.
In addition to serial mice and bus mice, there are "keyboard mice". On PS/2s
there are two identical-looking keyboard ports, labeled (with icons) "mouse" &
"keyboard". Both are 8 or 9 pin mini-DIN's that look like the regular PC
keyboard port only smaller. I don't know what logical protocol the keyboard
mouse speaks. Physically, the connector is eventually connected to the
keyboard processor (often an 8042). The same keyboard processor that decodes
the keyboard decodes the mouse. PS/2s have this port, many newer ISA/EISA
motherboards do as well.
All things considered, UNIX users are probably best off going with a serial
mouse (most current clone motherbords give you two serial ports, so you can
dedicate one to this and still have one for the all-important modem). Not only
are the compatibility issues less daunting, but a serial mouse loads the
multitasking system less due to interrupt frequency. Beware that most clone
vendors, being DOS oriented, bundle M-type mice for which UNIX support is
presently spotty, and they may not work with your X. Ignore the adspeak about
dpi and pick a mouse/trackball that feels good to your hand.
On the other hand: PS/2 mice deliver quadrature output (raw mouse output that
all mice speak) straight to the computer. This is also how Atari and Amiga
mice work. This is quite nice, because it makes the mouse simpler (and
therefore more reliable), and because you only get interrupts when the mouse is
actually doing something. This also means that if your PS/2 mouse breaks you
can get a cheap Atari or Amiga mouse (and they *are* cheaper) to replace it
without sacrificing mechanical quality (which is the important part).
VII. When, Where and How to Buy
If you're a serious UNIX hacker for either fun or profit, you're probably in
the market for what the mail-order vendors think of as a high-end or even
`server' configuration, and you're going to pay a bit more than the DOS
lemmings. On the other hand, prices keep dropping, so there's a temptation to
wait indefinately to buy. A tactic that makes a lot of sense in this market,
if you have the leisure, is to fix in your mind a configuration and a trigger
price that's just a little sweeter than the market now offers and buy when
that's reached.
Direct-mail buying makes a lot of sense today for anyone with more technical
savvy than J. Random Luser in a suit. Even from no-name mail-order houses,
parts and system quality tend to be high and consistent, so conventional
dealerships don't really have much more to offer than a warm fuzzy feeling.
Furthermore, competition has become so intense that even mail-order vendors
today have to offer not just lower prices than ever before but warranty and
support policies of a depth that would have seemed incredible a few years back.
For example, many bundle a year of on-site hardware support with their medium-
and high-end "business" configurations for a very low premium over the bare
hardware.
Note, however, that assembling a system yourself out of mail-order parts is
*not* likely to save you money over dealing with the mail-order systems
houses. You can't buy parts at the volume they do; the discounts they command
are bigger than the premiums reflected in their prices. The lack of any
system warranty or support can also be a problem even if you're expert enough
to do the integration yourself --- because you also assume all the risk of
defective parts and integration problems.
Cruise through "Computer Shopper" and similar monthly ad compendia. Even if
you decide to go with a conventional dealer, this will tell you what *their*
premiums look like.
If you do a lot of PC buying, you will want to subscribe to ClariNet
Communications's "Street Price Report". This is issued every other Thursday;
you can have it emailed to you, or get it from an FTP site and decrypt it using
an emailed key. It's basically a digest of current prices for all kinds of PC
hardware and software, collected from the ad sections of major magazines
including "Computer Shopper" and "PC Magazine". Once you've cruised the
magazines, you know what you want and are after the lowest price, you can nail
it without fail with the Street Price Report.
Another alternative to conventional dealerships (with their designer "looks",
stone-ignorant sales staff, and high overheads that *you* pay for) is to go
with one of the thousands of the hole-in-the-wall stores run by immigrants from
the other side of the International Date Line. They're usually less ignorant
and have much lower overheads; they do for you locally what a mail-order house
would, that is assemble and test parts they get for you from another tier of
suppliers. You won't get plush carpeting or a firm handshake from a white guy
with too many teeth and an expensive watch, but then you didn't really want to
pay for those anyway, right?
A lot of vendors bundle DOS 5.0 and variable amounts of DOS apps with their
hardware. You can tell them to lose all this cruft and they'll shave $50 or
$100 off the system price. However, David Wexelblat observes "there are at
least two situations in which the Unix user will need DOS available: 1) most,
if not all, EISA configuration utilities run under DOS, and 2) SCSICNTL.EXE by
Roy Neese is a godsend for dealing with SCSI devices on Adaptec boards."
Don't forget that (most places) you can avoid sales tax by buying from an
out-of-state mail-order outfit, and save yourself 6-8% depending on where you
live. If you live near a state line, buying from a local outfit you can often
win, quite legally, by having the stuff shipped to a friend or relative just
over it. Best of all is a buddy with a state-registered dealer number; these
aren't very hard to get and confer not just exemption from sales tax but
(often) whopping discounts from the vendors. Hand him a dollar afterwards to
make it legal.
(Note: I have been advised that you shouldn't try the latter tactic in
Florida -- they are notoriously tough on "resale license" holders).
(Note II: The Supreme Court recently ruled that states may not tax out-of-state
businesses under existing law, but left the way open for Congress to pass
enabling legislation. Let's hope the mail-order industry has good lobbyists.)
On the other hand, one good argument for buying locally is that you may have to
pay return postage if you ship the system back. On a big, heavy system, this
can make up the difference from the savings on sales tax.
VIII. Questions You Should Always Ask Your Vendor
A. Minimum Warranty Provisions
The weakest guarantee you should settle for in the mail-order market should
include:
* 72-hour burn-in to avoid that sudden infant death syndrome. (Also,
try to find out if they do a power-cycling test and how many repeats
they do; this stresses the hardware much more than steady burn-in.)
* 30 day money-back guarantee. Watch out for fine print that weakens this
with a restocking fee or limits it with exclusions.
* 1 year parts and labor guarantee (some vendors give 2 years).
* 1 year of 800 number tech support (many vendors give lifetime support).
Additionally, many vendors offer a year of on-site service free. You should
find out who they contract the service to. Also be sure the free service
coverage area includes your site; some unscrupulous vendors weasel their way
out with "some locations pay extra", which translates roughly to "through the
nose if you're further away than our parking lot".
If you're buying store-front, find out what they'll guarantee beyond the
above. If the answer is "nothing", go somewhere else.
B. Documentation
Ask your potential suppliers what kind and volume of documentation they supply
with your hardware. You should get, at minimum, operations manuals for the
motherboard and each card or peripheral; also an IRQ list, and a bad-block
listing if your Winchester is ESDI rather than IDE or SCSI (the latter two
types of drive do their own bad-block mapping internally). Skimpiness in this
area is a valuable clue that they may be using no-name parts from Upper
Baluchistan, which is not necessarily a red flag in itself but should prompt
you to ask more questions.
C. A System Quality Checklist
There are various cost-cutting tactics a vendor can use which bring down the
system's overall quality. Here are some good questions to ask:
* Is the memory zero-wait-state? One or more wait states allows the vendor to
use slower and cheaper memory but will slow down your actual memory subsystem
throughput. This is a particularly important question for the *cache*
memory!
* Is the monitor non-interlaced? Does it have a tilt-and-swivel base? Is it
*color*? Yes, if you don't see it in the ad, ask; some lowball outfits will
try to palm off so-called "black & white VGA" monitors on you. What's the
vertical scan rate? 60Hz is SVGA standard; 72Hz is VESA standard and
minimal for flicker-free operation; 80Hz is cutting-edge. What's the dot
pitch? .31mm is minimal, .28mm or .27mm is good. You need .28mm for X.
A slightly larger dot pitch is acceptable in a larger monitor (16" or more).
* If you're buying a factory-configured system, does it have FCC certification?
While it's not necessarily the case that a non-certified system is going
to spew a lot of radio-frequency interference, certification is legally
required --- and becoming more important as clock frequencies climb. Lack
of that sticker may indicate a fly-by-night vendor, or at least one in
danger of being raided and shut down!
IX. Things to Check when Buying Mail-Order
A. Tricks and Traps in Mail-Order Warranties
Reading mail-order warranties is an art in itself. A few tips:
Beware the deadly modifier "manufacturer's" on a warranty; this means you have
to go back to the equipment's original manufacturer in case of problems and
can't get satisfaction from the mail-order house. Also, manufacturer's
warranties run from the date *they* ship; by the time the mail-order house
assembles and ships your system, it may have run out!
Watch for the equally deadly "We do not guarantee compatibility". This gotcha
on a component vendor's ad means you may not be able to return, say, a video
card that fails to work with your motherboard.
Another dangerous phrase is "We reserve the right to substitute equivalent
items". This means that instead of getting the high-quality name-brand parts
advertised in the configuration you just ordered, you may get those no-name
parts from Upper Baluchistan --- theoretically equivalent according to the
spec sheets, but perhaps more likely to die the day after the warranty expires.
Substitution can be interpreted as "bait and switch", so most vendors are
scared of getting called on this. Very few will hold their position if you
press the matter.
Another red flag: "Only warranted in supported environments". This may mean
they won't honor a warranty on a non-DOS system at all, or it may mean they'll
insist on installing the UNIX on disk themselves.
One absolute show-stopper is the phrase "All sales are final". This means you
have *no* options if a part doesn't work. Avoid any company with this policy.
B. Special Questions to Ask Mail-Order Vendors Before Buying
* Does the vendor have the part or system presently in stock? Mail order
companies tend to run with very lean inventories; if they don't have your
item in stock, delivery may take longer. Possibly *much* longer.
* Does the vendor pay for shipping? What's the delivery wait?
* If you need to return your system, is there a restocking fee? and will the
vendor cover the return freight? Knowing the restocking fee can be
particularly important, as they make keep you from getting real satisfaction
on a bad major part. Avoid dealing with anyone who quotes more than a 15%
restocking fee --- and it's a good idea, if possible, to avoid any dealer
who charges a restocking fee at all.
C. Payment Method
It's a good idea to pay with AmEx or Visa or MasterCard; that way you can stop
payment if you get a lemon, and may benefit from a buyer-protection plan using
the credit card company's clout (not all cards offer buyer-protection plans,
and some that do have restrictions which may be applicable). However, watch
for phrases like "Credit card surcharges apply" or "All prices reflect 3% cash
discount" which mean you're going to get socked extra if you pay by card.
X. Which Clone Vendors to Talk To
I went through the March 1992 issue of Computer Shopper calling vendor 800
numbers with the following question: "Does your company have any
configurations aimed at the UNIX market; do you use UNIX in-house; do
you know of any of the current 386 or 486 ports running successfully
on your hardware?
I didn't call vendors who didn't advertise an 800 number. This was only partly
to avoid phone-bill hell; I figured that toll-free order & info numbers are so
standard in this industry sector that any outfit unable or unwilling to spring
for one probably couldn't meet the rest of the ante either. I also omitted
parts houses with token systems offerings and anybody who wasn't selling
desktops or towers with a 386/33DX or heavier processor inside.
After plundering Computer Shopper, I called up a couple of "name" outfits that
don't work direct-mail and got the same info from them.
The answers I get revealed that for most clone vendors UNIX is barely a blip on
the screen. Only a few have tested with an SVr4 port. Most seem barely aware
that the market exists. Many seem to rely on their motherboard vendors to tell
them what they're compatible, without actually testing whole systems. Since
most compatibility problems have to do with peripheral cards, this is a
problem.
Here's a summary of the most positive responses I got:
A --- Advertises UNIX compatibility.
C --- Has known UNIX customers.
I --- Uses UNIX in-house.
T --- Have formally tested UNIX versions on their hardware.
F --- Have 486/50 systems
* --- Sounded to me like they might actually have a clue about the UNIX market.
Vendor A C I T F * Ports known to work
--------------- - - - - - - -----------------------------------------------
ARC . . X X . . SCO XENIX 2.3.2, SCO UNIX 3.2.1
AST . X X X X * SCO UNIX 3.2.4, ODT 2.0 Microport V/4
Allegro . . X X . . SCO XENIX 3.2.4
Altec . X . X . . XENIX (no version given).
Ares . X X X X * AT&T 3.2, ISC (version unknown)
Basic Time . X X X X * SCO XENIX 2.3.2, have in-house UNIX experts.
Binary Tech . X . X X . Claims to work with all versions.
Blue Dolphin . X . X X * SCO XENIX.
CCSI X X . . X . They've used SCO XENIX, no version given.
CIN . X . . . . SCO UNIX (version not specified)
CSS . X . X . * SCO 3.2.2, ISC 3.0, SCO ODT. See Will Harper.
Centrix X . . . . . No specifics on versions.
Compudyne . X X X X . Couldn't get details on which versions.
Comtrade . X . X X . Couldn't get details on which versions.
Datom X X X X X . SCO XENIX 3.2.
Dell X X X X X * See Dell SVr4 data.
Desert Sands X X . X X . SCO UNIX 3.2.4
Digitech . X . X . . SCO UNIX 3.2.1, XENIX 2.3.1
EPS X X X X . * SCO XENIX 3.2.4, ISC & AT&T (versions not sp.)
Gateway 2000 X X X X X * SCO UNIX 3.2.0. XENIX 2.3.4 ISC 3.0, ESIX 4.0.3
HD Computer . X . X X . SCO UNIX 3.2, SCO XENIX 3.2.2
HiQ . X . X . . SCO UNIX (version not specified)
Infiniti . X . X X . SCO UNIX (versions not specified)
Insight . . X . X . SCO XENIX 3.2.4. No tech support for UNIX
Keydata X . X X X * SCO version 4, ISC 3.2
Legatech . X . . X . SCO UNIX, ISC (versions not specified)
MicroGeneration . . X . . . Uses XENIX.
MicroLab X . . . . . SCO UNIX, SCO XENIX
MicroSmart X X . X . . SCO XENIX (version not specified)
Microlink X . . X X . SCO XENIX (version not specified)
Myoda X X . X X . SCO XENIX 3.2.2, ISC 3.2
Naga . X . X X * SCO & XENIX 3.2.
Northgate X X . X X * SCO UNIX 3.2
PC Brand . X X X . . SCO XENIX, ISC UNIX
PC Professional . X . X . . ISC 3.2
PC-USA X X . X . . ISC 5.3.2 and SCO 3.2
Profex . X . X . . SCO XENIX 3.2.
Royal Computer . X . . X . No details on versions.
SAI X X . X X . SCO UNIX 3.2.2.
Santronics . . X X X . SCO XENIX 3.2.4
Solidtech . X . . . . Dell (no version given), ISC 3.2.
Strobe . . . X X . SCO, Microport, ISC (no version numbers given)
Swan X X X X X * SCO 2.3.1, UNIX 3.2, ISC 3.2v2.0.2
TriStar . X X X X * SCO UNIX 3.2.2, XENIX 2.3.2, ISCr4
Zenon . X . X X * SCO UNIX (version not specified)
Zeos . X X X X * SCO XENIX 3.2.4, AT&T 3.2
Special notes about a few vendors who appear to have a clue:
Ares targets some of its systems for UNIX CAD use. They have a house wizard
name Ken Cooper (everybody calls him "K.C.").
EPS targets some 486 EISA configurations for UNIX.
Swan doesn't know the UNIX market very well yet, but their project manager
wants a bigger piece of it and is interested in doing some of the right
things. They have a house wizard, one John Buckwalter.
Dell, of course, supports an industry-leading SVr4 port. They're a bit on
the pricy side, but high quality and very reliable. Lots of UNIX expertise
there; some of it hangs out on the net.
Zeos is on the net as zeos.com, with a uunet connection; they host a UNIX BBS.
They have an in-house UNIX group reachable at supp...@zeos.com; talk to Ken
Germann for details. There are biz.zeos.general and biz.zeos.announce groups
on USENET.
Special notes about a lot of vendors who appear to have *no* clue:
Vendors where I couldn't get a real person on the line, either because
no one answered the main number or because I couldn't raise anyone at
tech support after being directed there: Sunnytech, Quantex, AMS, USA
Flex, Lapine, Syntax Computer, MicroTough, PAC International, The Portable
Warehouse.
Vendors where the question met with blank incomprehension, puzzlement,
consternation, or "We've never tested with UNIX": Allur, AmtA, Aplus, HiTech,
Locus Digital Products, LodeStar, TriStar Computers, Ultra-Comp, UTI Computers,
PC Turbo Corp, Evertek, Microcomputer Concepts, Jinco Computers, UWE, ToughCom,
System Dynamics Group, Terribly Fast Bus Systems.
Vendors who understood the questions but had no answer: Bulldog Computer
Products, LT Plus, Standard Computer, JCC.
Vendors who said "Yes, we're UNIX-compatible" but had no details of any tests:
CompuCity.
Vendors who said "Go ask our motherboard vendor": Ariel Design, Lucky Computer
Co., V-com, Professional Computer, MicroLine, MileHi.
Vendors who sent me to a toll number: Absec, Hokkins, New Technologies, Mirage.
Vendors that believe they have UNIX customers, but can't be any definite than
that: Austin Computer Systems, PC Professional, Treasure Chest Computer
Systems, CompuAdd Express, FastMicro, MidWest Micro.
Final note:
If you order from these guys, be sure to tell them you're a UNIX customer
and don't need the bundled DOS. This will shave some bucks off the system
price, *and* it may encourage them to pay more attention to the UNIX market.
--
Send your feedback to: Eric Raymond = e...@snark.thyrsus.com
Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!darwin.sura.net!jvnc.net!
netnews.upenn.edu!dsinc!bagate!cbmvax!snark!esr
From: e...@snark.thyrsus.com (Eric S. Raymond)
Newsgroups: comp.unix.sysv386,comp.unix.bsd,comp.os.mach,news.answers
Subject: PC-clone UNIX Software Buyer's Guide
Summary: A buyer's guide to UNIX versions for PC-clone hardware
Message-ID: <1jjQdy#6dbp5J1VKBgV42Tvsg0glhbB=esr@snark.thyrsus.com>
Date: 7 Dec 92 19:44:34 GMT
Expires: 7 Mar 93 00:00:00 GMT
Sender: e...@snark.thyrsus.com (Eric S. Raymond)
Followup-To: comp.unix.sysv386
Lines: 2516
Approved: news-answers-requ...@MIT.Edu
Archive-name: pc-unix/software
Last-update: Mon Dec 7 14:40:22 1992
Version: 9.0
You say you want cutting-edge hacking tools without having to mortgage the
wife'n'kids? You say arrogant workstation vendors are getting you down? You
say you crave fast UNIX on cheap hardware, but you don't know how to go about
getting it? Well, pull up a chair and take the load off yer feet, bunky,
because...
This is the PC-clone UNIX Software Buyer's Guide posting,
current to Dec 7 1992.
What's new in this issue:
* New BSD/386 info.
* More Consensys bugs.
Gentle Reader: if you end up buying something based on information from this
Guide, please do yourself and the net a favor; make a point of telling the
vendor "Eric's FAQ sent me" or some equivalent. The idea isn't to hype me
personally, I've already got all the notoriety I need from doing things like
_The_New_Hacker's_Dictionary_ --- but if we can show vendors that the Guide
influences a lot of purchasing decisions, I can be a more powerful advocate for
the net's interests, and for you.
0. CONTENTS
I. INTRODUCTION. What this posting is. How to help improve it. Summary of
the 386/486 UNIX market, including 6 SVr4 products, SCO UNIX (an SVr3.2), and 2
BSD ports. What's new in this issue.
II. GENERAL CONSIDERATIONS. A brief discussion of general hardware
requirements and compatibility considerations in the base SVR4 code from UNIX
Systems Laboratories (referred to below as the USL code). None of this
automatically applies to SCO or the two BSD-like versions, which break out the
corresponding information into their separate vendor reports.
III. FEATURE COMPARISON. A feature table which gives basic price & feature
info and summarizes differences between the versions.
IV. VENDOR REPORTS. Detailed descriptions of the major versions and
vendors, including information collected from the net on bugs, supported
and unsupported hardware and the like.
V. UPCOMING PORTS, FREEWARE VERSIONS, AND CLONES. Less-detailed descriptions
of other products in the market.
VI. HARDWARE COMPATIBILITY TABLES. A set of tables summarizes vendor claims
and user reports on hardware compatibility.
VII. FREEWARE ACCESS FOR SVR4 SYSTEMS. Information on the SVR4 binaries
archive.
VIII. FREE ADVICE TO VENDORS. Your humble editor's soapbox. An open letter
to the UNIX vendors designed to get them all hustling to improve their products
and services as fast as possible.
IX. ACKNOWLEDGEMENTS AND ENVOI. Credit where credit is due. Some praises
and pans. What comes next....
Note: versions 1.0 through 4.0 of this posting had a different archive name
(386-buyers-faq) and included the following now separate FAQs as sections.
pc-unix/hardware -- (formerly HOT TIPS FOR HARDWARE BUYERS) Useful general
tips for anybody buying clone hardware for a UNIX system. Overview of the
market. Technical points. When, where, and how to buy.
usl-bugs -- (formerly KNOWN BUGS IN THE USL CODE). A discussion of bugs
known or believed to be generic to the USL code, with indications as to which
porting houses have fixed them. None of this applies to the two BSD-based
versions.
Readers may also find material of interest in Dick Dunn's general 386 UNIX
FAQ list, posted monthly to comp.unix.sysv386 and news.answers.
I. INTRODUCTION
The purpose of this posting is to pool public knowledge and USENET feedback
about all leading-edge versions of UNIX for commodity 386 and 486 hardware. It
also includes extensive information on how to buy cheap clone hardware to
support your UNIX.
This document is maintained and periodically updated as a service to the net by
Eric S. Raymond <e...@snark.thyrsus.com>, who began it for the very best
self-interested reason that he was in the market and didn't believe in plonking
down several grand without doing his homework first (no, I don't get paid for
this, though I have had a bunch of free software and hardware dumped on me as a
result of it!). Corrections, updates, and all pertinent information are
welcomed at that address.
This posting is periodically broadcast to the USENET group comp.unix.sysv386
and to a list of vendor addresses. If you are a vendor representative, please
check the feature chart and vendor report to make sure the information on your
company is current and correct. If it is not, please email me a correction
ASAP. If you are a knowledgeable user of any of these products, please send me
a precis of your experiences for the improvement of the feedback sections.
At time of writing, here are the major products in this category:
Consensys UNIX Version 1.3 abbreviated as "Cons" below
Dell UNIX Issue 2.2 abbreviated as "Dell" below
ESIX System V Release 4.0.4 abbreviated as "Esix" below
Micro Station Technology SVr4 UNIX abbreviated as "MST" below
Microport System V Release 4.0 version 4 abbreviated as "uPort" below
UHC Version 3.6 abbreviated as "UHC" below
SCO Open Desktop 2.1 abbreviated as "ODT" below
BSD/386 (0.3 beta) abbreviated as "BSDI" below
Mach386 abbreviated as "Mach" below
The first six of these are ports of USL's System V Release 4. Until last year
there was a seventh, by Interactive Systems Corporation. That product was
canned after half of ISC was bought by SunSoft, evidently to clear the decks
for Solaris 2.0 (a SunOS port for the 386 to be released late in 1992). The
only Interactive UNIX one can buy at present is an SVr3.2 port which I consider
uninteresting because it's no longer cutting-edge; I have ignored it.
Consensys has an early version of Destiny (SVr4.2) out, but I don't have full
information yet. It's said to be considerably improved over their 4.0.3
product.
Earlier issues ignored SCO because (a) 3.2 isn't leading-edge any more and (b)
their `Version 4' is a 3.2 sailing under false colors. Can you say deceptive
advertising? Can you say bait-and-switch? Can you say total marketroid-puke?
However, the clamor from netters wanting it included was deafening. The day
SCO landed an unsolicited free copy of ODT on my doorstep I gave in. I don't
expect to actually use it, but I summarize the relevant facts along with
everything else below. Note that ODT is their full system with networking and
X windows; what they call SCO UNIX is missing most of those trimmings.
BSD/386 is *not* based on USL code, but on the CSRG NET2 distribution tape.
Complete sources are included with every system shipped! Mach386 is basically
BSD tools with the monolithic Mach 2.5 kernel and does entail a USL license;
it's based on the Tahoe BSD distribution. For a few extra bucks, you can
get Mach 3.0 (a true microkernel) with *source*!.
AT&T's own 386 UNIX offering is not covered here because it is available and
supported for AT&T hardware only.
All the vendors listed offer a 30-day money-back guarantee, but they'll be
sticky about it except where there's an insuperable hardware compatibility
problem or you trip over a serious bug. One (UHC) charges a 25% restocking fee
on returns. BSDI offers a 60-day guarantee starting from the date of receipt
by the customer and says: "If a customer is dissatisfied with the product, BSDI
unconditionally refunds the purchase price." Dell says "30 day money-back
guarantee, no questions asked".
Some other ports are listed in section V.
II. GENERAL CONSIDERATIONS
To run any of these systems, you need at least the following: 4 MB of RAM and
80MB of hard disk (SCO says 8MB minimum for ODT 2.0; Dell 2.1 also requires 8
MB minimum). However, this is an absolute minimum; you'll want at least 8 MB
of RAM for reasonable performance. And depending on options installed, the OS
will eat from 40 to 120 meg of the disk, so you'll want at least 200 meg for
real work. To run X you'll need a VGA monitor and card, and 12-16MB RAM would
be a good idea.
Installation from these systems requires that you boot from a hi-density floppy
(either 3.5" or 5.25"). Most vendors offer the bulk of the system on a QIC 150
1/4-inch tape; otherwise you may be stuck with loading over 60 diskettes! BSDI
offers the distribution not only on QIC-150 tape but also on CD-ROM. They'll
even sell you a CD-ROM reader for US$225 (or you buy the same Mitsumi drive at
Radio Shack or Best Buy for US$199+tax). In general, if the initial boot gets
far enough to display a request for the first disk or tape load, you're in good
shape.
USL SVr4 conforms to the following software standards: ANSI X3.159-1989 C,
POSIX 1003.1, SVID 3rd edition, FIPS 151-1, XPG3, and System V Release 4 ABI.
4.0.4 ports conform to the iBCS-2 binary standard. The SVr4 C compiler (C
Issue 5) includes some non-ANSI extensions (however, note that as of mid-1992,
no SVr4 ports other than AT&T's have been formally POSIX-certified).
SCO conforms to the following standards: ANSI X3.159-1989 C, POSIX 1003.1 FIPS
151-1, XPG3, System V Release 3 ABI, and SVID 2nd Edition. Despite the
marketing droids hacking at its version number, SCO is not conformant to System
V Release 4 or SVID 3rd Edition.
All SVr4 versions include support for BSD-style file systems with 255-character
segment names and fragment allocation. In general this is a Good Thing, but
some SVr3.2 and XENIX binaries can be confused by the different size of
the inode index. You need to run these on an AT&T-style file system. SCO
UNIX 3.2v4 (thus, ODT 2.0 but not 1.1) has an `EAFS' file system which adds
symlinks and long filenames. Old SCO binaries can be confused by long
filenames.
All SVr4 versions include the UNIX manual pages on-line. Dell stocks
Prentice-Hall's SVr4 books and will sell them to you with your system (in lieu
of printed manuals) at extra cost. You can order them direct from
Prentice-Hall at (201)-767-5937. Warning: they ain't cheap! Buying the whole
13 volumes will cost you a couple hundred bucks. Esix, Microport and UHC
have their own manual sets derived from the same AT&T source tapes as the
Prentice-Hall set; Esix charges extra for them, but Microport and UHC both
include them with their systems.
SVr4 includes hooks for a DOS bridge that allows you to run DOS applications
under UNIX (the two products that actually do this are DOS Merge and VP/ix).
Most vendors do not include either of these with the base system, however.
All these systems support up to 1024x768 by 256 color super-VGA under X. The
640x480 by 16 colors of standard VGA is no problem; everybody supports that
compatibly. However, X servers older than the Roell or X11R5 version (that is,
MIT X11R4 or anything previous) are hard to configure for the clock timings of
your controller and monitor scan frequency unless you have one of the standard
combinations USL supports or your vendor has configured for it.
There are a couple of known hardware compatibility problems the USL code
doesn't yet address. See the KNOWN BUGS section near the end of this document.
III. FEATURE COMPARISON
To interpret the table below, bear in mind the following things:
All these products except BSDI/386, Mach386 and SCO ODT are based on the
SVr4 kernel from UNIX Systems Laboratories (USL), an AT&T spinoff. Thus they
share over 90% of their code and features. Product differentiation is done
primarily through support policy, bug-fix quality and add-on software.
The `USL support?' column refers to the fact that USL support is a separate
charge from the source license. With the former, a porting house gets access
to AT&T's own OS support people and their bug fix database, and the porting
house's bug fixes can get folded back into the USL code.
These systems come either in a "crippled" version that supports at most two
simultaneous users, or an unlimited version. Generally the vendors do allow
you to upgrade your license via a patch disk if your requirements, but this
invariably costs slightly more than the base price difference between 2-user
and unlimited systems.
The "run-time" system in the price tables below is a minimum installation,
just enough to run binaries. The "complete" system includes every software
option offered by the vendor; it does *not* bundle in the cost of the
Prentice-Hall docs offered by some vendors as an option. You may well get
away with less, especially if you're willing to do your own X installation.
The `Upgrade plan' section refers only to upgrades from previous versions
of the same vendor's software.
The numbers under support-with-purchase are days counted from date
of shipment. The intent is to help you get initially up and running.
The engineer counts below are as supplied by vendors; .5 of an engineer
means someone is officially working half-time. The `Uses USENET' column is
`yes' if there is allegedly at least one person in the engineering department
who reads USENET technical groups regularly and is authorized to respond to
USENET postings reporting problems.
The `DOS Bridge' row gives the version number of DOSMerge supplied with the
system, if any. DosMerge 2.0 has roughly the caoabilities of DOS 3.0, though
it is reported to be quite flaky and hard to configure. DOSMerge 2.2 has the
capabilities of DOS 5.0.
A dash `-' means the given feature or configuration is not offered. A `yes'
means it is currently offered; `soon' means the vendor has represented that it
will be offered in the near future. A `no' means it's not offered, but there's
some related information in the attached footnote.
Vendor SCO Cons Dell Esix MST uPort UHC BSDI Mach386
Base version: 3.2.2 4.0.3 4.0.4 4.0.4 4.0.3 4.0.4 4.0.3 BSD Mach
USL support? ?? ?? y y n y ??(a) n n
System price:
Run-time
2-user 595 - - 384 249 500 695 - -
Unlimited 1295 - - 784 449 1,000 1,090 - -
Complete
2-user 3090 995 995(b) - 799 3,000 1,990 - -
Unlimited 4290 1195 1295(b) 1,607 999 3,500 2,385 995(c) 995(d)
Printed docs? y(f) - y(e) y(e) - y(f) y - -
Upgrade plan?
From SVr3.2 y - y y - y - - -
Future SVr4s ?? - (h) (g) - (h) - - -
Support
W/purchase: 30 (i) 90 (j) 30 30 30 60 30
800 number? y - y - - - - - -
By contract y n(k) y n(j) y y y y y
Support BBS? y y - y - y soon - y
FTP server? y - y y - - - y y
Read USENET? y ?? y y - y n(l) y y
# Engineers:
Support: 60+ 1 5 2 2 4 2 1.5 1
Development: 55+ ??(m) 10 ~20 3 6 27 6.5 5
Distribution media:
3.5" 1.44MB y y(n) - y y y y - y
5.25" 1.2MB y y(n) - y y y y - y
60MB ctape y y - y y y y - -
125MB ctape - - - - y - y - -
150MB ctape - - - - y y y y -
250MB ctape - - y - - - - - -
2GB DAT - - y - - - - - -
Via network? - - y - - - - - -
X options:
X11/NeWS R3 - - - y - - y - -
MIT X11R4 y(o) - y - - - - - y
AT&T Xwin 3 - - - - y - - - -
AT&T Xwin 4 - - - y - y - - -
Roell X386 - - - - y - y - -
X11R5 - y y - - - - y soon
Open Look - - 4i 1.0 2.0 4i 4i - -
Motif 1.1.4 1.1 1.1.4 1.1.0 1.1.2 1.1.3 1.1.3 (p) soon
X.desktop 3.0 - - - - 3.0 2.0 - -
Also included:
DOS bridge? 2.2 - 2.2 - - soon - soon -
SLIP? y - y y - y soon y y(q)
PPP? y ?? - n(r) - - soon soon n
(a) UHC had a support contract at one time but may have let it lapse. I
expect to have better information on this soon.
(b) This price is for customer-installed UNIX. If it's factory-installed on
Dell hardware, it's $500 less.
(c) $995 is for credit-card CDROM orders; POs are $50 more; QIC-150 is $50
more. Educational site licenses are available for $2K each.
(d) Previous issues alleged that "No unlimited licenses have been sold yet."
Feedback from the net indicates that all MtXinu systems now being sold
are unlimited.
(e) Extra-cost option.
(f) With complete system only.
(g) Small media charge. Note: if you upgrade from a 2-user to multi-user
ESIX, you pay full price.
(h) Free with support contract, charge otherwise (charge ~$500).
(i) 90 days or until product is installed successfully.
(j) Unlimited free phone support.
(k) Charges by the half-hour phone call.
(l) UHC says they used to be net-active and want to be again when they can
afford the man-hours.
(m) Consensys explicitly refuses to release this information.
(n) There's an $80 media charge for the diskettes equivalent to the normal
60MB distribution tape.
(o) SCO's own X11R4 implementation.
(p) Motif for BSDI is available from a third party.
(q) At present, you must buy Mach386 Autosupport to get SLIP.
(r) Mark Boucher <m...@cam.org> has written a PPP driver for ESIX
The SCO information is included by popular demand for comparison purposes.
In the price figures, the `runtime' system is SCO UNIX 3.2v4; the `complete'
system is ODT with development tools.
In general, the SVr4 market breaks into two tiers. The bottom tier is
Consensys and MST; low-ball outfits selling stock USL with minimal support for
real cheap. The top tier is Dell, Esix, Microport and UHC; these guys are
selling support and significant enhancements and charge varying premiums for
it. Your first, most basic buying decision has to be which tier best serves
your needs.
One further note: it *is* possible to buy some of these systems at less than
the list the vendor charges! I found some really substantial discounts in one
mail-order catalog ("The Programmer's Shop"; call 1-(800)-421-8006 to get on
their mailing list, but be prepared to wade through a lot of DOS cruft).
IV. VENDOR REPORTS
Vendor reports start here. Each one is led by a form feed.
NAME:
SCO Open DeskTop
VENDOR:
The Santa Cruz Operation, Inc.
400 Encinal Street
PO Box 1900
Santa Cruz,CA 95061-1900
1-(800)-SCO-UNIX (sales)
1-(800)-347-4381 (customer service and tech support)
i...@sco.com --- product info by email, sales requests
supp...@sco.com --- support requests (support contract customers only)
SOFTWARE OPTIONS:
SCO's package and option structure is (excessively) complicated. At the
moment the `bundles' to keep track of are:
SCO UNIX System V/386 Release 3.2 Version 4.0
SCO UNIX networking bundle, consisting of:
SCO UNIX System V/386 Release 3.2 Version 4.0
SCO TCP/IP 1.2.0
SCO NFS 1.2.0
SCO Open Desktop 2.0:
SCO UNIX System V/386 Release 3.2 Version 4.0
SCO TCP/IP 1.2.0
SCO NFS 1.2.0
LAN Manager Client, PC-NFS daemon, PC-Interface server
X (X11R4 server/clients, Motif 1.1.4, X.desktop 3.0)
DOS Merge (2.2)
Note that Ingres (the database) has been removed from the ODT bundle since 1.1.
There is a special Ingres price for ODT customers, and Ingres has committed
to offering a 50% discount till the end of '92.
ADD-ONS:
There are piles of them. I was most impressed by the docs for the CodeView
debugger and MASM assembler, but the presence of ISAM support would probably be
more significant to the ordinary commercial user.
SCO bundles with X also include 18 clients (what in marketingese are called
``personal productivity and groupware accessories and controls'') which
include: mail, help, edit, paint, term, print, login, clock, color, session,
mouse, lock, and admin (official names all prepended with "SCO") as well as
DOS, load, and calculator clients.
SUPPORT:
You get 30 days of free phone support with purchase.
ODT support is $895 per year.
SCO has BBS coverage and a local support operation in the UK as well as the
US; BBS coverage only Germany. Local support is, in theory, to be provided by
distributors.
FUTURE PLANS:
IPX/SPX (Novell networking support) will be added soon.
HARDWARE COMPATIBILITY:
See the appendix for details. SCO provides a Hardware Compatibility
Guide with its software.
COMMENTS:
The docs are impressive; you could get a hernia trying to lift them all.
TECHNICAL NOTES:
There's an `MPX' kernel available from SCO that supports multiprocessing.
Though this is a 3.2 kernel, SCO has added support for SVr4-like symbolic
links and long filenames to Version 4.
SCO has a standard driver announcement protocol which allows the
utility hwconfig(C) to print out detailed configuration info on hardware
attached to the machine.
SCO's cross-development and DOS emulation support is unusually rich. It
includes lots of system utilities for I/O with a DOS filesystem, as well as
cross-development libraries and tools in the Development System. Microsoft
Windows 3.0 and Windows 3.0 applications are supported (in real mode), and
future releases will support Windows 3.1 and associated applications.
Graphical MS-DOS applications are supported in CGA graphics mode within an X
window, and VGA graphics are supported in full-screen mode.
KNOWN BUGS
SCO tar(1) chokes horribly on long filenames and symbolic links.
This is scheduled to be fixed in the next maintenance supplement, MSv4.2.
WHAT THE USERS SAY:
XENIX is the UNIX port hackers love to hate, but at 70% of the market SCO
must be doing something right. In general, SCO UNIX and XENIX are reputed to
be a very polished and stable systems. Unfortunately, they also drive
developers crazy because of numerous tiny and undocumented divergences between
the SCO way and the USL-based releases.
REVIEWER'S IMPRESSIONS:
The SCO support system is heavily bureaucratized and prone to thrash when
processing questions of unusual depth or scope. While probably adequate for
the random business luser, hackers are likely to find the contortions
required to get to a master-level developer very frustrating.
SCO in general has the fairly serious case of corporatitis you'd predict
from their relatively large size --- no-comment policies and
compartmentalization out the wazoo.
On the other hand, they sent me an unsolicited free copy, and I got huge
amounts of useful technical and hardware-compatibility info "unofficially" from
SCOer Bela Lubkin <be...@sco.com>. Gee. Maybe I should flame vendors more
often... :-)
NAME:
Consensys UNIX Version 1.3
VENDOR:
Consensys
1301 Pat Booker Road
Universal City, TX 78148
(800)-387-8951
{dmentor,dciem}!askov!root
Note: Consensys is now shipping an early version of Destiny (SVr4.2) out, but
I don't have full information yet. It's said to be considerably improved over
this 4.0.3 product. If any Consensys user out there would care to help me
update this entry for the 4.2 product, I'd much apperciate it! Consensys
doesn't like me any more...
SOFTWARE OPTIONS:
None.
ADD-ONS:
Basically this is a stock USL system with the stock USL bugs, except the
installation sequence has been improved considerably. Good tools for
configuration management and system administration on a network of Consensys
machines are included.
SUPPORT:
You get free phone support until your system is installed, to a maximum of
90 days. After that they charge per half-hour of phone time. They like to
do support by fax and callback.
They have 1 (one) support tech. Ask for Reuben.
They have a support BBS at (416)-752-2084.
Knowledgeable customers report they're good about supporting the bits they
wrote (see below) but terrible at dealing with generic SVr4 problems.
FUTURE PLANS:
They haven't settled on an upgrade policy yet.
There are plans for a disk array product.
HARDWARE COMPATIBILITY:
See the appendix for details.
Though most reports say the Consensys PowerPorts board is fine for UUCP use,
at least two USENETters have reported problems with interactive sessions; see
below.
TECHNICAL NOTES:
The X stuff is straight off the MIT X11R5 tape, patchlevel 8.
KNOWN BUGS:
This port probably uses the stock USL 4.0.3 libraries. Thus it probably
has the known bug with sigvec() and may have the rumored bug in the BSD-
compatibility string functions.
One `Andy', mailing from <hoop...@ucsu.Colorado.EDU> says "You should also
blast Consensys for advertising that they provide DOS file system utilities.
They do, but they were written for DOS 2.0! They do NOT work for DOS 5.0..."
Syd Weinstein <s...@dsinc.dsi.com> reports: "The most major [bug in the
PowerPorts support] is delays in various output codes.... Even if not using
the multi-screen stuff, a clear to end of line escape code, and some others,
cause noticable delays in the output. (About 0.1 seconds). It makes running
Elm a real bitch". He is in touch with Consensys about this.
It has been reported on USENET (by Gerry Swetsky <lis...@vpnet.chi.il.us>
among others) that if you drop off of a PowerPorts line without manually
closing all your sessions, the unclosed sessions may be accessible to the next
person to pick up the line.
Gil Kloepfer, Jr. <g...@limbic.ssdl.com>, managing the Houston UNIX User's
Group's system, says that during interactive use the board frequently does not
handle typeahead properly (this may be related to Syd Weinstein's problems with
EOL delay). He also says he hasn't been able to bring up stable UUCP with the
board.
Jim Bray <b...@wcuvax1.wcu.ed> adds " Let me throw in a few more specifics.
In version 1.3,
/usr/lib/tapecntl is a symlink to itself (it actually existed in 1.2),
and
/usr/sadm/sysadm/menu/backup_service/extended
/usr/sadm/sysadm/menu/restores/extended
are both symlinks to their parent dirs.
I was able to get nfs and rfs working with version 1.2, altho' it
was done by sort of shaking things around until they worked for no
clear reason. I have just finished removing and reinstalling them to
try yet again to get the 1.3 versions working, but whether I try the
v4net crap or the sysadm crap, neither of them work now. In general,
removing and reinstalling network packages never works. It might work
if I reinstalled the whole system, but I think a Linux or 386BSD
useable on my Gateway 486/EISA/SCSI systems will come along fairly
soon.
Once anything has gone wrong with the nfs/rfs stuff, it seems to be
fixable only by a system reload. When I originally brought 1.3 up on
one of our machines, the other was still running 1.2. BOth rfs and nfs
had been working. The 1.3 v4net code does a lot of remote-execing, and
it attempted to rexec things that exist in 1.3 but not 1.2: in other
words: 1.3 is apparently not backwards-compatible. This probably made
a hopeless mess of the nfs/rfs databases at that point."
COMMENTS:
Their UNIX product is an outgrowth of their main line of business, selling
serial boards. It is easy to configure the OS to support the board.
WHAT THE USERS SAY:
I've spoken with one experienced wizard using Consensys and seen a detailed
email report from another. They're happy, although they both warn that newbies
should probably *not* try this at home :-).
On the other hand, Consensys has a dismal reputation on USENET; horror
stories of nonexistent followup on bugs abound. They'll need to work hard to
shuck their take-the-money-and-run image. Better followup on the reported
serial-port board bugs would be a big help. Unfortunately, Consensys's favored
response seems to be to deny that they have a problem.
One customer (J.J. Strybosch, <j...@ubitrex.mb.ca>) reports that Consensys
charged his credit card for more than they quoted him. If you deal with them,
watch your credit card statement carefully.
REVIEWER'S IMPRESSIONS:
These guys have the toughest support policy of any vendor and
obviously don't want to hear from you once you've gotten past initial
boot.
A Consensys marketroid that I spoke with twice while gathering this
information offered to send me an evaluation copy of their system. They were
clearly hoping for some good publicity if I like it. However, I doubt they
like me that well any more...
Consensys explicitly refuses to say how many development engineers they have
on staff. In this and some other matters (such as the way they deal with
allegations of PowerPorts problems) they've adopted a corporate style that
appears defensive, evasive, secretive, and not conducive to trust. I couldn't
make their V.P. of sales understand that this appearance is a serious
liability in dealing with UNIX techies and distinguishes them from the
competition in a distinctly negative way.
NAME:
Dell UNIX System V Release 4 Issue 2.2.
VENDOR:
Dell Computer
9505 Arboretum Road
Austin TX 78759
(800)-BUY-DELL (info & orders)
(800)-624-9896 (tech support: x6915 to go straight to UNIX support)
i...@dell.com --- basic Dell info
supp...@dell.com --- support queries
SOFTWARE OPTIONS:
Basically, there aren't any. You get the development system with all the
trimmings for a lower list than anybody else in the top tier. Whaddya
want, egg in yer beer?
ADD-ONS:
Dell bundles a DOS bridge (Locus 2.2, supporting DOS 5.0) with their base
system. They also include cnews, mmdf, perl, elm, bison, gcc, emacs gdb, Tex,
network time protocol support, and other freeware, including a bunch of nifty X
clients! Also included: the Xylogics Annex server for TCP/IP network access.
FrameMaker is also included, but runs in demo mode only until you buy a
license token from Unidirect.
SUPPORT:
Dell *does* support their UNIX on non-Dell hardware. They are quite
definite about this. They will deal with software problems reported from
non-Dell hardware, but you're on your own when dealing with hardware
incompatibility problems unless you can reproduce the problem on a
Dell PC. However, it is also policy that if you lend them the offending
hardware, they will work with the vendor to come up with a fix.
You get 90 days of free phone support on a toll-free number, starting on
resceipt of your registration card (no card, no support). Yearly service
contracts range are $350 per year for the limited license, $500 for the
unlimited.
There are 6 engineers in their first line and 4 in their second-line support
pool.
Dell accepts software problem reports from anyone, Dell or non-Dell
hardware and whether or not they have a support contract. If you don't have
a support contract, don't count on getting a reply acknowledging the report.
Dell maintains a pair of Internet servers (dell1.dell.com and
dell2.dell.com) which hold patches, updates and free software usable with
Dell UNIX.
About upgrades, Dell says "If you have a support contract, the upgrade is
free, unless we've added something with significant royalty burden to us. We
may make a charge at that point. We didn't when we added Graphical Services
4.0 at the introduction of Dell UNIX 2.1. If you don't have a contract, then
the cost is basically Media+Royalty+Admin+Shipping."
FUTURE PLANS:
X.desktop 3.0 will be supported soon. NeWS isn't going to happen at all;
they couldn't get it to work reliability.
Dell has demonstrated a 486 port of NeXTSTEP at trade shows.
Dell is going to move to Solaris someday. However, policy is that they're
not going to phase out SVr4 until at least a year after their first *reliable*
version of Solaris, in order to provide an upgrade path.
TECHNICAL NOTES:
The big plus in the Dell code is that they've fixed a lot of the annoying
bugs and glitches present in the stock USL tape.
The installation procedure has been improved and simplified. You can
install Dell UNIX through your network from another Dell box once you've booted
the hardware with a special disk provided.
Both benchmarks and anecdotal reports make them significantly faster than a
stock USL system. Interestingly, Dell's manager for UNIX development tells me
this is all due to bug fixes and careful choices of some OS parameters.
A source at Dell has asked me to point out that Dell's SLIP can be
set up, configured, and stopped while UNIX is running; some other
versions (such as SCO's) require a reboot. However, others claim that
SCO's can actually be reconfigured without a reboot and that the SCO
*manuals* are at fault here for misleading people.
Dell device drivers are *very* unlikely to work on other SVR4 versions.
Dell includes some kernel extensions (not required, so other SVR4 device
drivers should work) to make life in support a little easier. A program
called showcfg will list all recognised device drivers and the IRQ,
I/O address, shared memory and so on. The device driver has to register
this info. Dell has told USL how to do this, it's up to them when or even
if they want to use this in a future release.
Dell device drivers are also auto configuring, for the most part. Check out
/etc/conf/sdevice.d/* and see how most of the devices are enabled, but with
zeroes in all fields for IRQ, I/O and memory. Those are autoconfiguring
drivers. Dell thinks that this makes life much easier; you only need to set
one of the configurations that they probe for! The device registration helps
this, by eliminating possible overlapping memory or I/O address usage. (On the
other hand, idconfig(1) is no longer helpful, when I/O, IRQ and mem are all
zero). The 2.2 release adds a utility `setcfg' which can be used to remove
unneeded drivers, shrinking the kernel.
Dell UNIX also has drivers for the Dell SmartVu found on some machines (a
little four character LED display on the front panel). By default this shows
POST values, then disk accesses, finally "UNIX" when running and "DOWN" when
halted. You can write to the device.
Dell's SCSI tape driver includes ioctls to control whether hardware
compression is used.
Some Dell systems have a reset button. On the Laptops these are wired
directly to the CPU. On the desktop and floor-standing systems Dell UNIX can
catch the interrupt; it's used to do a graceful (init 0) shutdown. Other
UNIXes will do a processor reset when the button is pushed.
About 95% of 2.2 was built using GNU cc for a significant performance
improvement over pcc.
KNOWN BUGS:
Uucico fails when sending more than 12 files to another machine. Fixed
in 2.2; a patch is available free from Dell for earlier versions.
Performance monitoring of uucp transfers doesn't work. Creating
/var/spool/uucp/.Admin/perflog results in uucico logging statistics to the file
correctly. However, using uustat -tsysname results in either a memory error or
you just being returned to the shell with no output. This bug is known to
Dell and being worked on now.
Merge is seriously buggy in many areas. It takes ages to start up in an
xterm and then sometimes crashes in the process. Attempting to use its
simulated expanded memory results in the system becoming slowly corrupted which
later results in virtual terminals disappearing and the system gradually
locking up. Really fun stuff! And it can only cope with 1.44M discs. These
are generic Merge problems, not really Dell's but Locus's fault.
There are some dropped stitches in the supplied USENET tools. The nntp
server has been compiled for a dbm history file while c-news has been compiled
for dbz. With nntpd this only shows on the ARTICLE <message-id> command which
either returns that the article with that id can't be found or crashes the
server. Also, they forgot to include the nntpd manual page or nntpxfer. A
Dell source thinks these things have been corrected in 2.2.
Dell's device driver autoconfiguration doesn't properly set up the mouse
port on the ATI Graphics Ultra card. You need to either remove all other
mouse drivers or use the DOS install program to manually force the mouse IRQ
to 5.
HARDWARE COMPATIBILITY:
Dell doesn't maintain a list of non-Dell motherboards and systems known to
work. And they're not willing to talk about the list they don't maintain,
because it would amount to endorsing someone else's hardware.
Dell promises that you can bring its UNIX up on any Dell desktop or tower
featuring a 386SX or up (it's hard to get the product on to the notebooks).
Notebooks can't drive a QIC tape and there aren't drivers for the pocket
Ethernet or token-ring adapter.
Jeffrey James Persch <using a friend's account> reports that he couldn't
get the X supplied with Dell UNIX 2.1 to work with a Microsoft bus mouse hooked
to the mouse port on a Compaq 486/33M or Systempro.
Andrew Michael <Andrew.Mich...@brunel.ac.uk> says "If you're buying Dell
UNIX for non-Dell hardware, first try booting the Dell floppy on it. From
experience, some BIOS ROMs cause Dell SVR4 to lock up at the point where it
tries to talk to the hard disk. If it gets to the point where it asks you
whether you want to install or not you can be pretty sure that all is well. An
AMI or Phoenix BIOS is OK; be careful of anything else."
See the appendix for more.
COMMENTS:
Dell sells hardware, too :-). They are, in fact, one of the most successful
clonemakers, and will cheerfully sell you a Dell computer with SVr4 pre-
installed. Their systems are expensive by cloner standards (with as much as a
$1000 premium over rock-bottom street prices) but they have a rep for quality
and reliability their competition would probably kill for.
You can get Dell product information by sending an email request to
i...@dell.com.
WHAT THE USERS SAY:
Most people who've seen or used it seem to think pretty highly of the
Dell product, in spite of minor problems.
A user in England observes: "Dell is the only firm that I found supplying
Unix at the real monetary exchange rate, not the usual computer pounds=dollars
nonsense. In the UK the 2 user version costs 699 pounds, which is pretty close
to the US price in dollars. For those of us who don't live on the left-hand
side of the pond (there are a few of us!) that's a distinct advantage." He
adds "Dell's UK support is pretty good. Not as good as Sun, but then you don't
pay as much! From previous experience, SCO support in the UK is, well, pretty
non-existent."
REVIEWER'S IMPRESSIONS:
Right now, I'd have to call Dell the market leader in SVr4s. The
combination of low price, highest added value in features, and reputation for
quality makes them very hard to beat.
The only serious negative I've seen is that their support system seems to be
very badly overloaded, so you can end up on hold for a while when calling. The
techs themselves are sufficiently cranked about this that they'll complain of
understaffing and corporate shortsightedness on the phone to a stranger.
Dell has recently doubled their support staff and fixed a bad bug in their
call-handling system that was freezing the queue for up to two hours at a
time. This will certainly help matters.
On the other hand, Dell's UNIX development manager responded to the first
issue of this FAQ with about three hundred lines of intelligent, thoughtful and
extremely candid comment, including a whole pile of hardware-compatibility info
and a number of excellent suggestions for improving the FAQ. He has
continued to send voluminous, factual feedback to later issues --- an example
other UNIX vendors would do well to emulate!
NAME:
ESIX System V Release 4.0.4
VENDOR
Esix Computers
1923 E. St. Andrew Place
Santa Ana, CA 92705
(714)-259-3020 (tech support is (714)-259-3000)
supp...@esix.everex.com
ADD-ONS:
None.
SOFTWARE OPTIONS:
ESIX can be bought in the following pieces:
Unlim 2-user
Base system 784 384
Base system + Networking 866 396
Development system 131 N/A
GUI module (X, Motif, Open Look, X.desktop) 610 380
Note that the base system without networking cannot be upgraded to the
base system with networking; you'd have to replace at full cost.
SUPPORT:
Purchase buys you unlimited free phone support. However, be warned that
there are only two engineers assigned to the job and they are swamped.
Esix offers a support BBS at (714)-259-3011 and 3013 (the 11 line
has a Trailblazer on it). They plan to bring up an Internet server in
the near future.
Patches are available via anonymous ftp to esix.everex.com.
FUTURE PLANS:
They don't plan to support DOS Merge because it's still horribly buggy.
Later in '92 they plan to release a multiprocessing UNIX.
TECHNICAL NOTES:
Stephen J. Friedl and D'Arcy Cain <da...@druid.uucp> have written a device
driver for Everex's STEP systems that can control the LED array on the front of
the box.
Relative to 4.0.3, 4.0.4 includes numerous bug fixes, a rewritten SCSI
driver, and better SCO binary compatibilty. The GUI package is significantly
different, changing from a home-grown ESIX implementation of X to a licenced
implementation of AT&T's xwin implementation (with ESIX support for additional
video cards added in.
HARDWARE COMPATIBILITY:
See the appendix for details. ESIX supports an unusually wide
range of peripherals.
They advertise support for the Textronix X terminal.
No one has reported any incompatibility horror stories yet.
KNOWN BUGS:
According to Esix, this port uses the stock USL 4.0.3 libraries. Thus it
must have the known bug with sigvec() and may have the rumored bug in the BSD-
compatibility string functions.
James D. Cronin <j...@tropel.gca.com> writes: When developing X applications
under Esix, watch out for mmap(2) failure. This is caused by an incorrect
version of mmap() defined in libX11.a and libX11.so. This bug existed in Esix
4.0.3, and continues in 4.0.4 and the recently shipped Xwindow bug fix it
(which seems to have more bugs than the original version). One workaround is
to remove the offending file, XSysV.o, from libX11.a and link with the Bstatic
option.
COMMENTS:
Another subsidiary of a clonemaker (Everex). They don't sell bundled
hardware/software packages yet.
Esix will sell you manuals troffed off the SVr4 source tapes for somewhat
less than the cost of the Prentice-Hall books. The content is almost identical
but the organixation into volumes a little different.
Unlimited free support sounds wonderful, and might be ESIX's strongest
selling point. However, ESIX users on the net have been heard to gripe that in
practice, you get the support you've paid for from Esix --- that is, none.
That isn't at all surprising given Esix's staffing level. If this guarantee is
to be more than a hollow promise, their technical support has to get more
depth.
Evan Leibovich <e...@telly.on.ca> is a long-time netter who makes his living
as a consultant and owns an Esix dealership. He says you can get ESIX at a
substantial discount from him or other dealers, also that dealers are supposed
to do first-line support for their customers (which he does, but admits other
dealers often fail to). Evan is obviously devoted to the product and probably
the right guy to email first if you think you'd be interested in it.
WHAT THE USERS SAY:
Ron Mackey <r...@dsiinc.com> writes "In general, we are pleased with ESIX.
We still have problems driving the serial ports at speeds greater than 9600
baud. We also still see occasional PANICs. These appear to be related to
problems with the virtual terminal manager." This may be the generic USL asy
problem again.
William W. Austin <uunet!baustin!bill> writes "The support from Esix seems
to be usable if (a) you are a hacker, (b) you know unix (sVr4 internals help a
lot), and (c) you get past the sales guy who answers the help line (Jeff
[Ellis] is *very* helpful). If I were a computer-semi-literate, commercial
user who only wanted his printer to work, etc., I might be up a creek for some
problems (no drivers for some boards, no support for mouse tablets, etc., but
that's what VARs are for). All in all, the support is at least a little better
than what I expected for free -- in many cases it is *far* better than the
support I got from $CO (is SCO really owned by Ebenezer Scrooge?)"
[Note: Jeff Ellis has since left.]
A longer appreciation from Ed Hall <edh...@rand.org>: "I had a problem with
the ESIX X server. I got through to technical support immediately, and was
promised a fix disk. The guy on the phone was actually able to chat with on of
the developers to check to see if the disk would solve the problem. The disk
came four days later."
"On the other hand," he continues, "I get the feeling that ESIX has only
made a mediocre effort to shake out the bugs before releasing their system-- or
even their fixes. For example, they `repaired' their X server, but the new
server only ran as root (it made some privileged calls to enable I/O
ports)--they quickly had to release a second update to fix this new problem.
They obviously fixed a lot of things in the new server, and performance is
improved quite a bit as well, but the stupid error they made in the first
"fixed" version should have been found with only the most minimal of testing."
"They've done some work on the serial driver, but there are still some
glitches (occasional dropped characters on a busy system at 38400bps, and a
real doozy of a problem--a system panic--when doing simultaneous opens and
ioctl's on a tty0xh and ttyM0xh device. This latter problem was due to my
using the M0xh and 0xh devices improperly, but panics are inexcusable. No idea
if this is a SYSVR4 problem or due to their fixes.)"
"So my impressions of them are mixed. Perhaps I just lucked out in geting
such rapid response on my support call, but I was impressed by it nonetheless.
On the other hand, their QA needs work..."
REVIEWER'S IMPRESSIONS:
The tech I spoke with at Esix seemed knowledgeable, bright, and very
committed to the product. Nevertheless, when I asked what he thought
distinguished ESIX from the competition, he had no answer.
This reinforced the feeling I got from the spec sheets that Esix has kind of
an also-ran mentality, with no market strategy or clear priority for improving
SVr4 that positions it against its competition. It doesn't have Dell's
steak-with-all-the-trimmings appeal, it's not pushing price like Consensys or
support quality like UHC or performance like Microport. (I'm told that
at one time, Everex was the price leader).
When I asked Esix's chief marketroid about this, he said that he thinks
ESIX's best asset is that the product isn't going to go away, and muttered
unkind things about the possibility that Dell would deep-six their SVr4 in
favor of Solaris 2.0. This does not a long-term strategy make.
Despite numerous "repositionings" since I wrote the first version of this
comment in May 1992, I've seen no reason to change any of the above.
NAME
MST UNIX
VENDOR:
Micro Station Technology, Inc.
1140 Kentwood Ave.
Cupertino, CA. 95014
(408)-253-3898
sa...@mst.com (product info & orders)
c...@mst.com (support)
ADD-ONS:
None.
SOFTWARE OPTIONS:
C Development System
Networking
X11R4 and X11R3
Motif
Open Look
SUPPORT:
30 days of support free with purchase.
1 year of fax/email support is $299, 1 year of phone support is $599.
FUTURE PLANS:
They expect to upgrade to Motif 1.2 and X11R5 Summer '92. No plans for
4.0.4 yet.
HARDWARE COMPATIBILITY:
They've promised to email me a list of hardware known to work, which
will appear in a future posting.
They decline to release information on hardware known *not* to work
for fear of offending vendors.
KNOWN BUGS:
This port probably uses the stock USL 4.0.3 libraries. Thus it probably
has the known bug with sigvec() and may have the rumored bug in the BSD-
compatibility string functions.
The DOS support is only 2.0-compatible (< 32-meg DOS partitions).
COMMENTS:
Another outfit offering stock USL real cheap. They were actually the first
to try this (in Fall '91) and were the price leader until Consensys blew past
them.
These guys really want to sell you preinstalled UNIX on their clone
hardware. Configurations range from $1349 to $5599 and look like pretty
good value.
WHAT THE USERS SAY:
I have one experience report from Ray Hill, <h...@ghola.nicolet.com>, who's
been running MST on a 486 for a month or so. He says it works; elm, cnews, and
trn are up, so standard UNIX sources compile up and work fine. His only
criticism is the relative skimpiness of the printed docs.
Harlan Stockman <hwst...@snll-arpagw.llnl.gov> writes "MST has been very
helpful at every step of the way; phone and e-mail support have been timely."
Geoffrey Leach <ge...@ibmpa.awdpa.ibm.com> warns that some of the files
(specifically, socket library headers) necessary to build X11R5 are bundled in
the networking option --- this may meen you have to buy it even if you don't
actually intend to network any machines.
REVIEWER'S IMPRESSIONS:
Anyone who's been to a hobbyist computer expo in the last five years knows
that the low-price clone-hardware market is full of small, hungry companies run
by immigrants, often family businesses. Their English is sometimes a little
shaky but (in my experience) they're honest and their product is good, and
their prices are *real* aggressive.
MST seems to be one of these outfits. Since Consensys ended their promo
MST is now the low-price leader in this market.
NAME:
Microport System V Release 4.0 version 4
VENDOR:
Microport, Inc.
108 Whispering Pines Drive
Scotts Valley, CA 95066
(800)-367-8649
sa...@mport.com (sales and product info)
supp...@mport.com (support)
SOFTWARE OPTIONS:
Networking (TCP/IP, NFS)
Software Development
User Graphics Module (X GUIs)
Graphics Development Module (X toolkits + man pages).
DOS Merge
ADD-ONS:
A few freeware utilities are included, notably kermit(1) and less(1).
They include a single-user copy of a program called `JSB MultiView'. It's a
character-oriented desktop program that front-ends conventional UNIX services
for character terminals and also provides a calendar service and
pop-up phone-book. It's something like a character-oriented X windows; each
on-screen window looks like a terminal to the application.
SUPPORT:
The base price includes printed docs. This is effectively the same content
as the Prentice-Hall SVr4 books; both are troffed off the SVr4 source tapes.
They have been very lightly edited for the Microport environment.
The base price includes 30 days or 1 year of phone support respectively
depending on whether you bought the base or complete system. Support is
said to be excellent for serious problems, not so good for minor ones (this
is understandable if one assumes their support staff is very good but
overworked, a hypothesis which is plausible on other evidence).
They have a support BBS at (408)-438-7270 or 438-7521. However, the level
of activity is low; one customer said (late February) that they hadn't put
anything useful on it in six months (Microport responds that they've
been too busy hammering on r4 to spend lots of energy on it).
FUTURE PLANS:
DOS Merge will be folded into the system soon.
Also working on improved performance for the Adaptec 1742 and other SCSI
controllers, expect that in May.
Microport believes they have a lead in multiprocessing SVr4 UNIX and intend
to push it.
File-system support for CD-ROMs is coming.
HARDWARE COMPATIBILITY:
See the appendix for details.
Math co-processors: Cyrix 20/25/33, Intel 80387 20/25/33, Weitek.
No one has reported any incompatibility horror stories yet. Bernoulli boxes
and Irwin tapes won't fly, but who cares.
TECHNICAL NOTES:
When I asked what differentiates Microport from the other SVr4 products,
the answer I got is "performance". The Microport people feel they've put
a lot of successful work into kernel tuning.
And, indeed, benchmarks from independent sources show that Microport's
fork(2) operation is quite fast. Other vendors show about 60 forks per
second on the AIM Technologies SUITE II benchmarks; Microport cranks 80.
This is the most dramatic performance difference the AIM tools reveal
among any of these products. Microport's other benchmark statistics
are closely comparable to those of its competitors.
Microport also offers a multiprocessing SVr4 which will run on the
Compaq SystemPro, the ALR PowerPro, the DEC 433MP, and the Chips &
Technologies Mpax system.
Microport has moved the socket headers and libraries necessary to build X
out of the networking option package into the development system, so you
don't have to buy an extra module to hack X.
KNOWN BUGS:
According to Microport, this port uses the stock USL 4.0.4 libraries. Thus
it must have the known bug with sigvec() and may have the rumored bug in the
BSD-compatibility string functions.
David Wexelblat reports that "Microport's enhanced asy driver does not work
correctly (or at all) for hardware flow control - you can't open the ttyXXh
devices under any circumstances. This was true in 3.1, and is still true in
4.1. The good news is that SAS (Streams-FAS) works fine for modems. But SAS
won't work with the AT&T serial mouse driver. So I've got asy on my mouse port
and SAS on the other one on my dumb-card. [...] Microport is still prone to
silly errors. The Motif development system, which is described in the release
notes as being included with the Motif runtime system in the 'complete'
package, is in fact missing from the tape. They have it available seperately,
but I had to call them to get it. The 'pixed' application for X.desktop 3.0 is
compiled with shared libraries that are not included with the release. Hence
it does not work. I had to call them about this, too."
COMMENTS:
These people sold a lot of shrink-wrapped UNIXes years ago before going
chapter 11. They're back, leaner and meaner (with a total staff of just
15).
Microport says it's primarily interested in the systems-integration market,
where customers are typically going to be volume buyers qualifying for deep
discounts. Thus, they're relatively undisturbed by the certainty that their
high price point is losing them sales to individuals.
WHAT THE USERS SAY:
I've received one good comprehensive experience report, largely favorable,
from David Wexelblat <d...@mtgzfs3.att.com>.
REVIEWER'S IMPRESSIONS:
Microport is a small, hungry outfit with a lot to prove; they've already
gone bust once (I was a customer at the time :-() and they haven't yet
demonstrated that they've got a better strategy this time out.
They're perhaps a mite too expensive for the support quality they can offer
with less than fifteen people, and kernel-tuning isn't going to win them a
following on hardware that every year swamps those tweaks with huge increases
in speed for constant dollars. It may be that they're counting on the
multiprocessor version to be their bread-and-butter product; there, at least,
they're offering something that is so far unique and promises performance
levels unattainable with conventional hardware.
And, like UHC, they have techies answering the phones and the techies have a
clue. This certainly improves them as a bet for wizards and developers. If
multiprocessing is important to you, and/or you're looking for a small outfit
where you can develop personal working relationships with the tech people who
matter, Microport might be a good way to go.
They've offered to send me a copy of their OS gratis for review and
evaluation purposes.
NAME:
UHC Version 3.6
VENDOR:
UHC Corp.
3600 S. Gessner
Suite 110
Houston, TX 77063
(713)-782-2700
supp...@uhc.com
SOFTWARE OPTIONS:
Networking package (TCP/IP).
X + Motif
X + Open Look
ADD-ONS:
None reported.
SUPPORT:
The base price includes printed docs. This is effectively the same content
as the Prentice-Hall SVr4 books; both are troffed off the SVr4 source tapes.
30 days free phone support with purchase.
All their engineers take tech-support calls for part of their day.
They have 2 doing it full-time. The product manager is a techie himself
and takes his share of calls.
A support contract costs $1195 for one year. This includes 75% off
on all upgrades.
They are in the process of bringing up a BBS with a window into their
bug report and fix/workaround database.
It was emphasized to me that UHC wants to be known for the quality of
their support, which they feel is the product's strongest differentiating
feature.
FUTURE PLANS:
X11R5 by mid-May or thereabouts. They have it running now but don't
consider it stable enough to ship.
HARDWARE COMPATIBILITY:
See the appendix for details.
The asy driver in version 2.0 won't talk to the NS16550AFN UART, which
is supposed to be pin-compatible with the standard 16450.
KNOWN BUGS:
This port probably uses the stock USL 4.0.3 libraries. Thus it probably
has the known bug with sigvec() and may have the rumored bug in the BSD-
compatibility string functions.
COMMENTS:
They claim that according to USL they have the largest installed base of
SVr4 customers, and to have been first to market with a shrink-wrapped
SVr4 (in 1990).
UHC also claims to have performed and maintained IBM's official UNIX port
for the MicroChannel machines.
A subsidiary of Anam, "a holding company with a diversified portfolio".
WHAT THE USERS SAY:
The only comment I've yet seen on the UHC OS was an extended description of
a successful installation by a satisfied netter. He made it sound like a good
solid product.
I have one absolutely incandescently glowing report on UHC support from a
developer named Steve Showalter <shw...@Texaco.COM>. He says: "We've been
running UHC's OS for about a year now...been EXTREMELY happy with it. The
support we receive is without a doubt, the finest we have received from any
vendor."
Duke Smith (c/o som...@watson.bm.com) writes: "Another absolutely
incandescently glowing report on UHC support: I called the Programmer's Shop
about UHC & wound up talking to UHC tech support to find out if the sucker
would run on my machine. The guy took considerable time to explain all the
different things that might be causing the problem, and emphasized that the
same hardware problems which were probably causing Consensys not to run would
also hose UHC. This led me to contact ALR tech support (also a glower) who took
all of 1-1/2 days (not including shipping) to do the necessary upgrades, on
warranty because apparently their ads that it will run Unix are covered by
warranty. The glowing thing about UHC is, the guy helped me get a competitor's
port working, and I told him he was gonna get in dutch with the marketroids and
his response was that maybe I would remember them the next time I or someone I
knew needed a system. He's right. I'll use Consensys until I can afford
something better for my own system (it's still better than DOS...), but from
now on my clients will get pointed toward UHC, not Consensys, whose
absent-parent attitude is going to keep them from ever becoming anything but
the destitute hacker's Unix vendor."
On the other hand, William G. Bunton <w...@succubus.tnt.com>: "So, I give a
thumbs up for the product. I give a thumbs down for the company, and it's
enough that I'm taking my future business elsewhere." He tells a horror story
about the 2.0 version involving a three-month runaround, a letter to their VP
of marketing, and lots of broken promises. Apparently UHC does sometimes drop
the ball.
REVIEWER'S IMPRESSIONS:
I found both the people I talked to friendly, candid, technically
knowledgeable, and willing to answer sticky questions. I came away with a very
positive impression of the outfit's operating style.
There are experienced UNIX developers who value dealing with a small,
responsive outfit where they can develop good working relationships with
individuals. UHC says it likes to sell to wizards and might be a good choice
for these people.
The second time I called (*after* I'd formed the above impressions) one of
their guys offered to trade me a copy of UHC UNIX with all the trimmings for an
autographed copy of _The_New_Hacker's_Dictionary_. So they have taste, too.
I'm too ethical to let this sway my evaluations, but not too ethical to take
the software... :-)
NAME:
BSD/386
VENDOR:
Berkeley Software Design, Inc.
3110 Fairview Park Drive, Suite 580
Falls Church, VA 22042 USA
(800)-800-4BSD
bsdi-i...@bsdi.com
SOFTWARE OPTIONS:
None. You get an unlimited user license, binaries *and sources* for the
entire system (this includes X11R5 and full BSD networking sources with both
Internet and GOSIP OSI protocol stacks). What more could you want?
SUPPORT:
The purchase price include 60 days of phone support.
A telephone-support contract is $595 per year; email-only support is
$295/year; upgrade only is $185/year.
FUTURE PLANS:
Capability to run SVr3.2 binaries (including SCO binaries) in 1993.
They intend to add a DOS bridge by the end of '92.
The current release (0.3) is a fairly stable beta. Rob Kolstad sez:
"Our current release (November 30, 1992) is titled Gamma 4 for
legal reasons. Our 1Q1993 release will be big-fixes for even
better quality."
HARDWARE COMPATIBILITY:
See the appendix for details. New drivers are being added all the time.
Most multiport serial boards aren't supported (they're working on it).
BSD/386 supports the RISCOM/8 multiport serial card (SDL: 508-559-9005) and
includes a driver for the MAXPEA serial cards.
Rob Kolstad says BSDI has been very pleased with the cooperation
they've received from systems vendor Technology Power Enterprises. He
says: "In a world of commodity products, they differentiate themselves
by good service. When we (as operating system developers) have any
problems with their boxes, they're happy to help us out in finding and
fix problems -- even when the problem is hardware!" Dave Ingalz of
that company has developed a BSD/386-ready configuration for people who
might wish to buy one; call 510-623-3834.
TECHNICAL NOTES:
Alone among the 386 UNIX versions described here, this version is *not*
based even in part on USL code and has no AT&T license restrictions. Rather,
it derives from Berkeley UNIX (the CSRG Networking 2 release, somewhere between
4.3 and 4.4).
Many of the BSD/386 tools, including the compiler, are GNU code.
This system's libraries, header files and utilities conform to X3J11, POSIX
1003.1 and POSIX 1003.2 standards. POSIX Certification is schedule for the
first half of 1993.
COMMENTS:
What these people are trying is audacious --- something functionally like
the SVr4 merge, but starting from a ported BSD kernel and with System V
compatibility hacks, rather than the other ways. By all accounts the product
is in far better shape right now than one would expect for a beta pre-release,
which argues that the developers have done something right.
WHAT THE USERS SAY:
The few who've seen this system display an evangelistic fervor about it.
REVIEWER'S IMPRESSIONS:
I expect this will become a hackers' favorite.
All this, and sources too...I salivate. I am tempted. Not sure I'm ready
to change OSs at the same time as I switch machines, though. SVr4's got better
continuity with the 3.2 I'm running now. Ghu, what a dilemma!
When I mentioned that I'm doing elisp maintenance for GNU EMACS these days,
Rob Kolstad, one of the principal developers, offered me a copy and a year
of support if I'd field their (so far nonexistent) EMACS problems.
NAME
Mach386
VENDOR:
Mt. Xinu
2560 Ninth Street
Berkeley, CA 94710
(510)-644-0146
mtxinu-m...@mtxinu.com
ADD-ONS:
Kernel sources! You get can sources for the Mach 3.0 microkernel for
$195 over base price.
SOFTWARE OPTIONS:
The base package includes: Mach 2.5 kernel and utilities, 4.3 BSD interface,
GNU utilities (GCC, GDB, GAS, EMACS, BISON), and on-line reference manuals (man
pages) for Mach and 4.3 BSD. The following options are available:
Networking (SUN NFS, TCP/IP networking from the Berkeley Tahoe release,
on-line NFS man pages).
X (X11R4 with programmer's environment and complete X manual pages).
On-line Documentation (Complete source for Mach and 4.3 documentation,
including Mach Supplementary Documents, System Manager's Documentation, 4.3 BSD
Programmer's Supplementary Documents, 4.3 BSD User's Supplementary Documents).
Optional Microkernel Add-on, Mach 3.0 (Complete Mach 3.0 microkernel source
code; complete build environment with tools to modify and rebuild the Mach 3.0
microkernel; binary BSD server which runs on top of the microkernel in place
of the standard /vmunix kernel; source for an example of a server (POE)
running on top of the Mach 3.0 microkernel and sources for some utilities
which are kernel-dependent.
SUPPORT:
You get 30 days phone support with purchase.
A support contract is available for $150 quarterly or $500 per year; this
includes upgrades. There is a support BBS open to contract holders only.
An ftp server at autosupport.mtxinu.com carries patches, enhancements and
freeware adapted for the system. That site also hists an NNTP server carrying
support newsgroups for MtXinu users. This service is called "auto-support". A
user writes: " They post bug reports/fixes, allow general user discussion, and
let registered users download updates. I have mixed feelings about
auto-support. The user activity on the news groups is pretty low, but Mt Xinu
responds to bug posts VERY quickly. Major updaes seem to occur about every 2-3
months. The cost is $150.00/quarter or $500/year. If you want the sources to
the 386-AT drivers and the build environment for the kernel, you need to buy an
auto-support subscription."
FUTURE PLANS:
They plan to move to OSF/1 this year. X11R5 and Motif support are
also in the works.
HARDWARE COMPATIBILITY:
See the Appendix for details.
Color X windows is supported on VGA boards via extended 8-bit color mode.
Toshiba and Toshiba-compatible floppy drives and controllers work.
All current motherboards tested have worked. There were a few problems with
early Compaq DeskPros. They add "Please note that we do not support the
microchannel bus, EISA extended modes, IBM PS2, and some NCR machines. We are,
however, considering new devices so let us know your interests!".
TECHNICAL NOTES
This product is essentially a 4.3 port built on the Mach project's
microkernel technology. This is a truly nifty architecture which builds a
4.3BSD-compatible kernel out of a collection of communicating lightweight
processes. The distinction between user and kernel mode almost vanishes, and
things like the schedular and virtual-memory manager which are normally
embedded deep in the kernel become semi-independent, modifiable modules.
COMMENTS:
Very appealing for the educational market --- lets CS students and hobbyists
tinker creatively with the guts of UNIX in a way that would be impossible under
more conventional UNIXes. It's not clear who else will be interested in this.
WHAT THE USERS SAY:
Eric Baur <e...@ventoux.assabet.com> writes:
"The system is a very faithful emulation of BSD43 on top of Mach. For our
purposes it is a super deal. For about $2000.00 in hardware and $995.00 in
software we have a Mach development platform that integrates almost seamlessly
into our network development environment. As a general-purpose UNIX (whatever
that means) Mach386 gives up a lot in features to the System V vendors.
(Virtual terminals, DOS emulation, etc etc) For the home hacker, it seems like
it would be a good deal. You obviously could never run "shrink-wrapped"
software, but most public domain and GNU stuff should port easily."
Mark Holden <l00...@eeyore.stcloud.msus.edu> adds "Mt. Xinu's tech support
is absolutely top-notch, and I've found them quite willing to deal with matters
even after the official support runs out. [...] Not that Mach386 is without
its quirks. I've had problems getting a Western Digital ethernet board to
work correctly, and things required a fair bit of tweaking to set things on a
smooth course, but then I've never worked with a BSD that didn't."
REVIEWER'S IMPRESSIONS:
Right now, this product is a solution looking for a problem --- a solution I
find technically fascinating, to be sure. But even the company admits to not
being sure who its market is. I wish 'em luck.
KNOWN BUGS:
Bugs reported in previous Guide issues with UUCP on bidirectional serial
lines have been fixed.
Eric Baur reports: "Fortunately, I got the micro-kernel add-on only as an
example for Mach 3.0 development. It is not nearly as stable as the mach 2.5
based production kernel. Our 486/33 EISA machine usually hangs within minutes
after booting the 3.0 kernel...Mt Xinu is completely up front about the limits
of the 3.0 stuff and is very helpful about trying to debug it."
V. UPCOMING PORTS, FREEWARE VERSIONS, AND CLONES.
There are three other commercial SVr4 UNIX ports on the market for which I do
not yet have detailed information. I hope to cover them in future issues.
They are covered in section VI, along with freeware UNIXes and UNIX clones.
PromoX UNIX:
This is said to be a bare-bones port by an outfit that mainly sells hardware.
The price quoted is $595 for a complete 2-user + devtools system.
PromoX Systems
1050 East Duane Avenue, Suite B
Sunnyvale, CA 94086
Tel: (408) 733-2966
Email: pro...@cup.portal.com
SORIX:
This is a SVR4 UNIX port enhanced for real-time work, offered by Siemens AG.
Siemens AG
AUT 189
Gleiwitzerstr. 555
8500 Nuremberg 1
Tel: 0911/895-2203
I don't yet know if this version is going to be sold in the US. In the info
I have, prices are quoted in Deutschmarks.
NeXTSTEP 486:
NeXT has a 486 port of the NeXT environment scheduled for beta release in
4th quarter '92.
There are some freeware alternative UNIXes available for the 386/486. None of
these are yet complete and mature hacking environments, but they show promise
(and require much less in minimum hardware to run). They are:
386BSD:
Under development by Bill & Lynne Jolitz & friends (this is the same 386BSD
project described in Dr. Dobbs' Journal some time back). This OS is based on
the NET/2 tape from Berkeley, strongly resembles the commercial BSD/386 release
described above, and like it is distributed with full source. The aim is to
produce a full POSIX-compliant freeware BSD UNIX. Version 0.1 is now out,
including FP emulation, SCSI support, coexistence with DOS, and many more new
features. Passwording has to be acquired separately due to US export
regulations, but the system is otherwise fairly complete; I have seen it run.
There's a lot of traffic in comp.unix.bsd about this project.
Linux:
This is a POSIX-emulating UNIX lookalike, being written from scratch and
currently in beta. At the moment, it's less complete than 386BSD because it
doesn't leverage as much pre-existing code, but the kernel and development
tools are up and usable. Linux is changing so fast that more description would
probably be more misleading than enlightening. There's an active linux group
on USENET, comp.os.linux, and a (now less active) linux-activists mailing list;
to subscribe, mail to "linux-activists-reque...@niksula.hut.fi". Up-to-the
minute info is also available by fingering torva...@kruuna.helsinki.fi.
Hurd:
This is the long-awaited and semi-mythical GNU kernel. It's being worked on
by the Free Software Foundation (the people who brought you emacs, gcc, gdb and
the rest of the GNU tool suite) but it's not ready for prime time yet. It's
said to be a set of processes layered over a Mach 3.0 kernel. The 386BSD and
Linux developments both lean heavily on GNU tools.
There is one other not-quite-freeware (cheapware?) product that deserves a
mention:
Minix:
This is a roughly V7-compatible UNIX clone for Intel boxes, sold
with source by Prentice-Hall for $169 (there's an associated book for
a few bucks more). It's really designed to run in 16-bit mode on 8086
and 286 machines, though the UK's MINIX center offers a 32-bit kernel.
UUCP and netnews clones are available as freeware but not supplied
with the base system. A large international community is involved in
improving Minix; see comp.os.minix on USENET for details.
These freeware and "cheapware" products exert valuable pressure on the
commercial vendors. Someday, they may even force AT&T to unlock source to stay
competitive...
Finally, there is a class of commercial UNIX clones that claim to emulate UNIX
or improve on it without being derived from AT&T source. The major products
of this kind for 80x86 machines seem to be Coherent, QNX and LynxOS. The
following information about these has been supplied by various USENETters:
COHERENT is a small-kernel UNIX-compatible multi-user, multi-tasking
development O/S for $99.95 that uses less than 14Mb of disk space, runs on most
286-386-486 CPU systems, has a 64k limit C compiler and over 230 UNIX commands
including text processing, program development, administrative and maintenance
functions. It resides on a partition separate from DOS and can access the DOS
file system with the DOS command. It has no network or Xwindows support, but
cnews and rn have been ported and it has its own newsgroup, comp.os.coherent.
It is fully documented with both a comprehensive 1200 page manual and an
on-line manual. Mark Williams Company provides excellent support including a
UUCP access BBS and has just announced Release 4.0, the 386 version of
COHERENT.
QNX is a POSIX-compliant microkernel OS with real-time capability, targeted
to mission critical, performance sensitive applications like factory
automation, process control, financial transaction processing, and
instrumentation. They claim an installed base of over 200K systems worldwide.
The microkernel is only 7K and implements a message-passing model; other pieces
can loaded in at runtime, supporting anything from a small real-time executive
up to a full multi-user time-sharing system (including transparent DOS
emulation supporting Windows 3.1 in protected mode). QNX networking supports
standard protocol suites, but uses very fast, lightweight protocols for
QNX-to-QNX node communications; QNX machines on a network can be treated for
most purposes as a single large multiprocessor, and the OS itself can be
distributed across multiple nodes. Here is contact information for the vendor:
Quantum Software Systems Quantum Software Systems
175 Terrence Matthews Cr. Westendstr.19 6000 Frankfurt
Kanata, Ontario K2M 1W8 am main 1
Canada Germany
voice: (613) 591-0931 x111 (voice) voice: 49 69 97546156
fax: (613) 591-3579 (fax) fax: 49 69 97546110
usenet: stua...@qnx.com
QNX support is offered via voice and FAX hotline and a BBS. There is also
a newsletter and an annual international users' conference.
LynxOS is a 386 UNIX specialized for real-time work, available from Lynx
Real-Time Systems Inc. of Los Gatos, California. It includes TCP/IP, NFS, X,
etc. Most of the development tools are GNU. The kernel is pre-emptable and
supports threads and dynamically-loaded device drivers.
VI. HARDWARE COMPATIBILITY TABLES
These tables summarize vendor claims and user reports on which hardware will
work with which port.
To save space in the tables below, we use the following *one-letter*
abbreviations for the OS ports:
S SCO UNIX version 3.2v4
C Consensys UNIX Version 1.2
D Dell UNIX Issue 2.1
E ESIX System V Release 4.0.4
M Micro Station Technology SVr4 UNIX
P Microport System V/4 version 4
U UHC Version 3.6
B BSD/386 (0.3 beta)
X Mach386
A `c' indicates that the hardware is claimed to work in vendor literature.
A `y' indicates that this has been verified by a user report.
A `.' indicates that whether this combination works is unknown.
An `n' indicates that the vendor advises that the combination won't work.
A `*' points you at footnote info.
A blank column indicates that I have received no vendor info on the
hardware category in question.
The following general caveats apply:
* All ports support EGA, VGA, CGA and monochrome text displays.
* All ports support generic ISA serial-port cards based on the 8250 or 16450
UART. According to the vendors, the asy drivers on Dell, Esix, Microport,
BSD/386 and Mach386 support the extended FIFO on the NS16550AFN UART chip.
Indeed, Dell tech support will tell you this feature was present in the
base USL code. UHC says its 2.0 drivers *don't* talk to 16550s but
says that will be fixed in March '92. A user reports that SCO has
supported the 16550 since 3.2.2.
* I have not bothered listing ordinary ST-506/IDE/RLL drives, though lists
of them are given in vendor literature. This is a very mature commodity
technology; anything you buy should work with one of the supported
controllers unless it's defective.
* Vendors' supported hardware lists are not models of clarity. Some iterms
may be listed under a couple of different names because I don't know that
they're actually the same beast. I have been very careful not to make
assumptions where I am ignorant; thus, some hardware may appear less
widely supported than it actually is.
* These tables are grossly incomplete.
Also, be aware that there is a fundamental design problem in the ISA
architecture that can cause 8-bit boards used in a system with 16-bit
boards to flake out even if they're actually compatible. Jeremy Chatfield
of Dell describes it this way:
"We've seen (and fixed) this with several card combinations. If you have an 8
bit card and a 16 bit card in the same address range, then the address decoding
on the ISA bus will find that the 128KB range includes a 16 bit card. It
therefore programs itself for 16 bit I/O. If you then do I/O with the 8 bit
card, every other data byte is garbage. You will also have a reboot problem,
because the 16 bit card usually starts in 8 bit mode and has to be switched to
16 bit mode. If the switch back to 8 bit mode is not made, and the address
range is the c0000-d0000 range, close to the VGA BIOS, the VGA BIOS accesses
are screwed, because they are performed in 16 bit mode because of the above PC
H/W architectural problem. We include a deinit sequence in all the 16 bit
device drivers that causes a shutdown to reset the accesses to the safer 8 bit
mode. Of course, after a panic, the machine still has boards set up in 16 bit
mode, so you might observe the problem then.
This affects *all* PC OS's. I have seen cases where DOS failed to reboot
because of the same nonsense (network card in 16 bit mode in same address
region as VGA BIOS). Clever programming can resolve in several ways."
All the SVr4 systems inherit support for a fairly wide range of hardware from
the base USL code (version 4.0.3 or 4.0.4). This includes:
* All PC disk controllers (ESDI, IDE, ST-506 in MFM and RLL formats).
* The Adaptec 1542B SCSI adapter. Note: you'll have to jumper your
SCSI devices to fixed IDs during installation on most of these.
* Western Digital's 8013EBT Ethernet card, and its equivalents
the WD8003 and WD8013. SVr4v4 adds the 3Com 3C503.
* VGA adapters in 640x480 by 16 color mode.
* "C" protocol serial mice like the Series 7 and Series 9 from Logitech and
the PC-3 mouse from Mouse Systems (however, we've had one report of an
ostensible PC-3 clone called the DFI200H not working). See the "HOT TIPS"
section for details.
SCO UNIXes from 3.2.2 up and ODT 1.1 also support all these devices.
If you can fill in any of the gaps, or convert a `c' to `y', send me email.
S C D E M P U B X Systems
-------------------------------------------------------------------------------
c . c . Acer (all 386/486 models)
. . . c ACCELL 486/33 ISA and 386/40 ISA
. c . . ADDA AD-428P-25, Portable 486/25, 486/33, AD-328D-25
. c c . ALR Business VEISA 386/33-101
c . . . ALR (all 386 and 486 models)
c . . . applicationDEC 316,316+,325,325C,333,425,433MP
c . . . Apricot LS, LS 386SX, XEN-S 386
c . . y Arche 486, Master 486/33
. . c . AST (models not specified)
. c c . AST Premium (models not specified)
c . . . AST Premium 386,386/33,486/25T*E*,486/33T*E*
c . c . AT&T 6386 machines
. . c . Compaq (models not specified)
c c c . Compaq DeskPro 386/33.
c . . . Compaq DeskPro 486s/20,486/25/486/33L,386/20,386/25
c . . . Compaq Portable III 386, SystemPro
c . . y Compaq SLT 386s/20
. . . y CompuAdd 320
c y y y y y y . CompuAdd Model 333
. . . y CompuAdd 320
c . . . DEC DS486, DECpc 433, DECpc 433T
c . . . DECstation 320,325,425
c y . c . Dell 310,325,325P,333P,316SX,316LT,320SX,320LT.
c y . c . Dell 433P,425E,433E,425TE,433TE,4xx[DS]E,486[DP]xx.
. . . y DynaMicro 486/33
c . . . EasyData 386 model 333
c . . . Epson Equity 386/20PC,386/25,386SX; Epson PC AX3,AX3/25
. c c . Everex (models not specified)
y . . . Everex 33,386/20,486,486/33
. c c . Gateway 2000 (models not specified)
. . c y y Gateway 2000 (486/33 ISA)
. . . y . Gateway 2000 486/25
c . . . Groupil Uniprocessor 25MHz Tower
c . . . GRiDCase 1530,1550SX
. . c c High Definition Systems 486/25 ISA and 386 SX/16 ISA
. y . . High Definition Systems 386/40 ISA
c . . . HP 486 Vectra series
c . . . IBC 486
c . . . ITT 486
y . . . Micro Way Number Smasher 486/33
c . . . Mitac 386, MC3100E-02, S500
c . . . Mitsuba 386
c . . . Mitsubishi PC-386
. . . y MORSE PAT 386PX 386/40
. . . y MORSE KP 386T 386/33
c . . . NCR 316,316SX,3386
c . . . NEC 386/20,486/25, BusinessMate and PowerMate
y . . . NEC 386/33 BusinessMate
c . . . Noble 386
c . . . Nokia Alfaskop System 10 m52, m54/55
c . . . Northgate 33
. c c . Northgate 386/33
. y . . . Northgate 486/33
c . . . Olivetti 386/486 machines
c . . . Olivetti XP-9
y . . . Packard-Bell 386x
c . . . PC Craft PCC 2400 386
c . . . Phillips 386, P3464 486
. c c . Primax (models not specified)
c . . . SNI 8800-50, 8810-50, PCD series
c . . . Schneider 386 25-340, 386SX System 70
c . . . Siemens Data Systems Model WX200
c . . . Starstation
. . . y Tandy 3000
c . . . Tandy 4000
y . . . Tatung Force 386x
c . . . Tatung Force TCS-8000 386, TCS-8600 386
. c c . Tangent (models not specified)
. y . . Tangent 386/25C
. c y . Tangent 433E (486/33 EISA)
. . . y Technology Advancement Group EISA 483/33
. c c . Televideo (models not specified)
c . . . Televideo 386/25
c . . . Texas Instruments System 1300
. . . y Texas Instruments System 80486/33Mhz
c . . . Toshiba T3100,T3200,T5100,T5200,T8500,T8600
. . . y TPE 486/33 & 486/50
. c c . Twinhead (models not specified)
y . . . Twinhead 800 (486/33)
. c c . Unisys (models not specified)
c . . . Unisys PW2 Series 800/16,800/20,800/25
c . . . Victor 386 25, V486T
c . . . Wang MX200, PC 380
c . . . Wyse 386
n . . . Wyse Decision 486/33 (intermittent crashes)
c . . . Zenith 386 and 486 machines
. . . y Zeos 486DX-50
S C D E M P U B X Motherboards
-------------------------------------------------------------------------------
c . . AGI
y . . . A.I.R. 486/33EL w 256K cache
. c . ALR
. c . AMAX
c . c . AMI (model not specified)
y c . AMI Enterprise II (33 & 50)
y . . Amptron AMD386/40
. . y Amptron ISA 486DX/33
. c . ARC
n . c . Cache Computer
. c . Chips & Technologies chipset
y . c . Chips & Technologies 33DX
c c . Club AT
. c . DataExport
y . c . Dell
. c n . DTK (model not specified)
y . n . DTK 386/33
. . c EISA Tech 80386SX MHz
y . . . Eteq 386
y n . . Eteq 486
. c . Free Technology (model not specified)
. . . y Free Technology 486/33 EISA board
y . . . Free Technology 486/50DX
y y . . Gigabyte GA-486US 33MHz 256K Cache
c . . y Intel 302 (386/25 + 387)
. . y Intel 403E (486/33 EISA)
. c . Microlab
c y c y c Micronics 386/25
c c y c y Micronics 486/33 ISA
y . . Micronics 486/33 EISA
. c . Mitac
. . . Modular Circuit Technology 386/SX 16Mhz
y . . . Motherboard Factory 386/40, 486/33 (Northgate's OEM)
. c . Mylex (model not specified)
c c . Mylex MI-386/20
y y y y . Mylex MAE486/33
y y . . NICE 486DX/50 EISA
y . c . OPTI 486
. c . Orchid
. c . PC-craft
y . . TMC Research Corporation PAT38PC 25/386,33/386
y . . TMC Research Corporation PAT38PX 33/386,40/386
Notes:
* These two tables probably way *understate* the compatibility of most ports.
Most ISA or EISA motherboards will work with all of them. However, Jeff
Coffler <coff...@jeck.amherst.nh.us> reports: "I couldn't get the Cache
Computer CPU board to work at all with Dell UNIX, even though they claimed
they work with SCO. Flaky, timing-related failures."
* Quote from Kolstad, "The external caches on the most advanced
boards are usually not tested well for UNIX-like applications. We
see problems occasionally that disappear when the caches are disabled.
Once reproducible, the vendors can usually repair the problem."
* A source at UHC describes the DTK boards as "dogshit" --- he says they
generate a lot of spurious interrupts that DOS is too cretinous to be
bothered by but which completely tank UNIX. He says DTK seems uninterested
in fixing the problem. Other correspondents confirm that this has been
going on for several years. Avoid these boards till further notice.
* Dave Johnson <d...@gradient.com> reports that since upgrading from a 386 to
an Eteq 486, they've had lots of UHC random panics due to page faults in
kernel mode. UHC is looking into this.
* Some of the cards marked `supported' for SCO require the AGA EFS (Advanced
Graphics Adapters Extended Feature Supplement). (EFS's may be downloaded
for free via UUCP or FTP'd from uunet, but there is a media charge if they
are ordered on physical media from SCO).
S C D E M P U B X Video Cards Max Res ChipSet
-------------------------------------------------------------------------------
. . c y * . . Appian Rendition GRX 1024x768x256 TIGA34010
c . c y * . . Appian Rendition II, IIXE 1024x768x256 TIGA34010
c . c . . . . Appian Rendition III 1280x1024x256 TIGA34020
. . . c . . ARC V-16 (Paradise) ???? ????
. . . c c . AT&T VDC 600 (Paradise clone) SVGA ????
c . . . . . AST motherboard video 1024x768x256 WD90C31
c . . . AST VGA Plus 800x600x16 WDC
c . y . c . c ATI Ultra 1024x768x256 Mach 8
c . y . c . c ATI Vantage 1024x768x256 Mach 8
c . c c n y ATI Wonder+ SVGA N Wonder
c . . . . . ATI Wonder XL 1024x768x256 ????
. . . c c . ATI (type not specified) ???? ????
. . y . . . . Boca SuperVGA 1024x768 ET4000
c . . . . . Chips 451 800x600x16 N C&T451
c . . . . . Chips 452 1024x768x16 N C&T452
c . . . . . Compaq Advanced VGA 640x480x256 N ????
c . . . . . Compaq Plasma 640x400x2 N non-VGA
c . . . . . Compaq LCD VGA 640x480x16 N ????
c . . . . . Compaq VGC 640x480x16 N ????
c . . . . . Compaq AG1024 1024x768x256 ????
c . . . . . Compaq QVision 1024x768x256 ????
. c . . . c Compuadd Hi-Rez card w/1meg 1024x768 ET4000
c . . . . . Cornerstone SinglePage 1008x768x2 ????
c . . . . . Cornerstone PC1280 1280x960x2 ????
c . . . . . Cornerstone DualPage 1600x1280x2 ????
c . . . . . Cornerstone DualPage 150 2048x1560x2 ????
c . . . . . DEC 433w 1280x1024x256 TMS34020
c . . . . . DEC motherboard video 1024x768x256 WD90C30
c . y . . . . Dell motherboard video 1024x768x256 WD90C31
. . y . . c . Dell VGA 1024x768 ????
. c y c c y c y Diamond SpeedStar 1024x768 ET4000
c . . . . c Diamond Stealth 1280x1024x16 S3
c c . . . c Eizo MD-B07, MD-B10, Extra/EM 1024x768 ET3000
. . . . . y ELSA WINNER 1280x1024 82C480
. . c . . . Everex ViewPoint VRAM SVGA+ ????
. . c . . . Everex ViewPoint True Color SVGA+ ????
. . c . . . Everex UltraGraphics II EV-236 1664x1200 mono
c . c c c c Genoa 5300/5400 superVGA SVGA N ????
c c . c . c Genoa 6000, 6400 SVGA N ????
c . . . . . Grid 1500 laptop 640x400x2 CGA-like
y . c c . . Hercules monographics display 720x348 mono
c . . . . . HP UltraVGA 1280x1024x16 S3?
c . . . . . IBM 8514/A 1024x768x256 8514/A
c . y . . . . IBM VGA VGA VGA
c . . . . . IBM XGA 1024x768x256 XGA
c . . . . . IBM XGA-2 1024x768x256 XGA-2
c . . . . . Imagraph ITX 1280x1024x256 TMS34020
c . . . . . Intel motherboard video 1024x768x256 WD90C3x
c . . . . . Matrox MWIN1280 1280x1024x256 N ????
c . . . . . Matrox PG-1281-CV 1024x768x256 ????
. . c . . . MaxLogic SVGA ????
. . . . c . . Microfield V-8 1280x1024 ????
c . . . . . Microfield I8 1024x768x256 ????
c . . . . . Miro Magic 1280x1024x256 N 82C48
. . . . * . . Mylex GXE (EISA) 1280x1024 TIGA34020
. . . . . y Nth Engine/150 1280x1024 82C480
c . . . . . Number Nine GXi 1280x1024x256 TMS34020
. . c . . . Oak Technology OTI-067 1024x768x256 ????
c . . . . . Oak Technologies Oak 077 1024x768x256 Oak 077
c . . . . . Olivetti EVC-1 (EISA) 1024x768x256 82c452
. c . . . c Optima Mega/1024 1024x768 ET4000
c . . . . . Orchid Designer SVGA ET3000
c . . . . c Orchid Fahrenheit 1280x1024x16 S3
c c y c c c Orchid ProDesigner 800x600 ET3000
c c y y y . y Orchid ProDesigner II/1024 1024x768 ET4000
. . * y . . Orchid ProDesigner IIs 1024x768 ET4000
c . . . . . y Paradise VGA Plus SVGA PVGA1A
. c c c c c Paradise VGA Professional SVGA PVGA1A
c c c . c . c Paradise VGA 1024 SVGA WD90C00
c . . . . . Paradise 8514/A SVGA+ ????
. . . . . y PixelWorksWhirlWin 1280x1024 82C480
c . . . . . QuadRAM QuadVGA SVGA ????
. . . c c . Qume Crystal 1024x768 T4000
c . . . . . Renaissance Rendition II 1024x768 TMS34020
y c y y c . c Sigma Legend 1024x768 ET4000
. . . c c . Sigma VGA/H ???? ????
c . c c c . STB EM-16 VGA, EM-16+ VGA SVGA ET3000
c . . . . . STB Extra-EM SVGA ET3000
. c c c . c STB PowerGraph w/1meg 1024x768 ET4000
. c . . . c Swan SVGA with VCO chip 1024x768 ET4000
c . c . . . Tecmar VGA AD SVGA ET3000
c . . . . . Toshiba Grid 1500 laptop 640x400x2 CGA-like
. c . . . c TRICOM Mega/1024 1024x768 ET4000
c . . c . . Trident SuperVGA ???? T880
c . . . . c Trident TVGA 8900 1024x768 T8900
. . . c c . Tseng Labs VGA 1024x768 T4000
. . c . . . Vectrix VX1024 (TI-34010) 1024x768 ????
c . . . . . Verticom MX/AT 800x600 ????
c . c c c . Video7 FastWrite VGA 800x600 x2, x16 ????
c . . c c . Video7 VRAM VGA 800x600x16 Video7
c . . c c . Video7 VRAM II VGA SVGA Video7
c . . c c . Video7 VEGA EGA 640x380 Video7
c . . . . . Video7 VGA1024i SVGA Video7
c . y . . . . Zenith/Bull motherboard video 1024x768x256 WD90C31
In this table, an `SVGA' resolution code signifies the following resolutions:
1024x768 at 2 and 16 colors, 800x600 at 2, 16, 256 colors, and 640x480 at 2,
16, 256 colors. SVGA+ adds 1280x1024 at 2 or 16 colors. Some non-interlace
boards are marked with N.
Caveats in interpreting the above table:
* All super-VGA cards will work at VGA resolutions and below (that is, resolu-
tions up to 640x480 in 16 colors).
* Because color is of secondary importance for most UNIX applications, I list
only the highest dot-density resolution of a board that supports more than
one. Some boards have lower resolutions with more colors.
* This list is not exclusive. Many (perhaps even most) dotted combinations
will work. UHC claims that any SVGA based on an ET3000, ET4000, Paradise
or Genoa chip-set will fly; Dell echoes this with regard to ET3000,
ET4000, WD90C0xx cards, and the same is probably true of all other vendors.
* The Renaissance GRX-II is the same board as the Appian Rendition II; the
company changed its name. The II/XE is compatible with the Rendition GRX
and the Appian Rendition II, it differs in architecture in that it supports
more DRAM and runs a little faster than the older cards. All
Rendition II type cards run at a maximum resolution of 1024x768-256,
the Renditon III runs at 1280x1024-256 with its full VRAM set.
* Consensys's list is just MIT's list of cards certified to work with X11R5;
Consensys is careful to note that they haven't tested all these themselves.
* An ESIX reseller says all the TIGA34010-based video cards are pretty much
alike and ESIX will drive any of them (the prudent user should probably ask
to see the card working before committing). ESIX also supports 720x348
resolution on cheap Hercules-compatible monochrome tubes, and the Everex
UltraGraphics display at 1664x1200 resolution.
* Beware the Trident and Oak chipsets. Many clone vendors bundle these with
their systems because they're cheap, but they break the Roell server and
some other X implementations. Also, they appear to argue with the WD8003EP
net card, and no re-arrangement of the jumpers seems to fix it.
* Third party server technology from companies like MetroLink can support
higher performance, higher resolution TIGA and proprietary technology.
* Dell's 2.2 adds X11R5 servers for VGA 640x480, 800x600 and for the Tseng
Labs ET4000 and WD90C11 in up to 1024x768 16 or 256 colour. Appian
Rendition II (formerly Renaissance) for 1024x768 TIGA 34010. Highest
performance from the ATI Ultra 1024x768 256 colour, and highest resolution
from the 1280x1024 256 colour JAWS (Dell proprietary card developed in
association with Lotus and MicroSoft)
* The Orchid ProDesigner IIs (top speed 80 MHz, not the 75MHz version) works
with both X386-1.2D and X386-1.2E (beta). It works ok with the ESIX 4.0.3
X11R4 stuff at any resolution under 1024x768. But the driver does *not*
work with 1024x768 (timings are way off). The vanilla ProDesigner II does
work correctly with both the X386 and the Esix X11's (R5 and R4,
respectively). Note: this info may change in ESIX 4.0.4, which uses a
different X.
* The Qume Crystal is a private-label version of the Tseng Labs VGA card.
S C D E M P U B X Mice
-------------------------------------------------------------------------------
c . y c y y y y (Logitech-compatible) 3-button serial mice (C protocol)
c . y c c n y (Logitech-compatible) 3-button bus mice (C protocol)
. . . c . n . ATI Wonder+ bus-mouse port
y . . . . c . ATI Graphics Ultra bus-mouse port
c . . . . . . HP C1413A Mouse
y . y . . . . IBM PS/2 keyboard mouse
c . y y c c n y Logitech MouseMan (M+ protocol)
c c y y c c c c . Logitech Trackman (serial, M+ protocol)
c c y c c n y Logitech Trackman (bus, M+ protocol)
c . . . . . . Logitech hi-res Keyboard Mouse
c . y c c c y Microsoft 2-button (serial, M protocol)
c . y c c n y Microsoft 2-button (bus, M protocol)
c . . . . . . Olivetti Bus Mouse
c . . . . . . Olivetti hi-res Keyboard Mouse
. . . . . . c SummaMouse
c . . . . . . Summagraphics Bitpad
Notes:
* See the discussion of mice at the beginning of this section for details.
* BSD/386 says it supports all 1200-9600 baud serial mice, specifying Logitech
as an example. This is probably true of all vendors.
* The MouseMan and TrackMan require a patch obtainable from SCO to run under
ODT 1.1; they're fully supported in 2.0.
* X11R5 (X386 1.2) supports all of the known mice on SVR4 in a native mode,
bypassing the mouse driver. This wasn't true with X11R4 (X386 1.1b).
So if you're using X386 1.2 exclusively, you can use (say) a MouseMan
regardless of which SVR4 you're using.
* Dell 2.2 includes an auto-configuring mouse driver that's supposed to
work with about anything. Non-factory-installed 2.2s may require a
patch from support to handle the Logitech Mouseman.
S C D E M P U B X Multi-port serial cards
-------------------------------------------------------------------------------
c . . . . . AMI lamb 4 and 8-port
. . y c c n Arnet (models not specified)
c . y . . . Arnet 2,4 and 8-port and TwinPort
c . . c c n AST 4-port
. . . . c n Central Data
c . . . c n Chase Research
c . c . c n Computone (models not specified)
c . y . . . Computone Intelliport
c . . . . . Computone ATvantage-X 8-port
c . . . . . Comtrol Hostess-4
c . . c c n Comtrol Hostess-8
. c . c y n Consensys PowerPorts
c . . . . . CTC Versanet 4AT and 8AT
c . y . . . Digiboard 4 and 8-port
. . y c c n Digiboard DigiChannel PC/8
. . . . . y . Digiboard Digichannel PC/Xe-16 (see note below)
y . y . y n Equinox
c . . . . . Kimtron Quartet 4-port
y . . c c c n Maxpeed
c . . . . . Olivetti RS232C Multiport board
c . . . . . Quadram QuadPort 1 and 5-port
. . . . . c . SDL RISCOM/8
y . y . c n Specialix
. . y . c n Stallion OnBoard
. . . . c n Stargate (models not specified)
c . . . . . Stargate OC4400 (4-port) and OC8000 (8-port)
c . . . . . Tandon Quad serial card
. . y . c n Technology Concepts
c . . . . . Unisys 4-port
Notes:
* Only SCO, Consensys, Dell, Esix and Microport listed multiport cards at all.
As some are `smart' cards which require special device drivers, you should
*not* assume that a board is supported on a particular port unless the
vendor explicitly says so.
* MtXinu says they have *no* multiport support right now.
* The Consensys PowerPort card has troubles; see the vendor report on
Consensys for details.
* The Chase, Computone, Intelliport and Specialix cards will run under
SCO using a vendor-supplied driver.
* The Maxpeed SS8-UX2 doesn't support RTS/CTS flow control, and requires
its own config scripts rather than using inittab and gettydefs. The
BSDI people think it works with their config stuff.
* Peter Wemm <Peter-W...@zeus.dialix.oz.au> writes: "In 2.1, Dell's drivers
(direct from Stallion) are flakey. I have been annoying the living daylights
out of the developers (Stallion) here in AUS, and their new drivers have an
`interaction' problem with the reboot mechanism in dell's kernel. A reboot
causes the VGA card to be disabled." Jeremy Chatfield of Dell replies:
"We haven't seen the problem he reports. Most likely the problem he's seeing
is an icky [generic] one for UNIX on a PC." He then proceeds to detail
the 8-16 clash described at the beginning of this section.
* Digiboard makes an SVr4 UNIX streams driver available via download for the
Digichannel PC/Xe-16.
S C D E M P U B X Disk controllers
-------------------------------------------------------------------------------
c . c c c . . Adaptec 2320/2322 (ESDI)
c . c . c . . Adaptec ACB 2730C (RLL)
c . y . c . . Adaptec ACB 2732C (RLL)
c . . . . . Compag 32-bit Intelligent Drive Array Controller
c . . . . . Compag 32-bit Intelligent Drive Array Expansion
. . c . c . c CCAT100A (IDE)
. . . c . . Chicony 101B
y . y c c . . Data Tech Corp 6280 (ESDI)
. . . c . c DTG 6282-24
. . c c c . . Everex EV-346 (ST506)
. . c c c . . Everex EV-348 (ESDI)
. . c c c . . Everex EV-8120 (IDE)
y . c . . . . Lark ESDI controller
. . c c c . . OMTI 8240 (ST506)
. . c . . c . PSI Caching controller (ESDI)
c . c . . . . SMS OMTI 8620 and 8627 (ESDI)
. . y . . c . Ultrastor 12C, 22F
y . y . c c c Ultrastor 12F
c . c . . n . Ultrastor 22C (caching EISA version of 12F)
. . y . c . . Ultrastor 22CA
c c y c c . . Western Digital 1003 (RLL)
c . . . . . Western Digital 1005
. . y . . . Western Digital 1006V-MM2 (ST506)
y . y y c . c Western Digital 1007 A,SE2 (ESDI)
c . . c . . Western Digital 1009 SE1/SE2
Notes:
* All these ports should support all standard PC hard-disk controllers (ESDI,
IDE,ST-506 in MFM and RLL formats).
S C D E M P U B X SCSI controllers
-------------------------------------------------------------------------------
c . . . . . . Adaptec 152x (non-bus mastering ISA host adapter)
y c y c y y c c y Adaptec 1540, 1542
c . n . . . . . Adaptec 1640 (MicroChannel version of 154x)
c . y c y c n c y Adaptec 1740,1742 (EISA) (1542 emulation mode)
c . . y . * c . Adaptec 1740,1742 (EISA) (enhanced mode)
. . . c . . . Always IN2000
y . c c . . . BusTek BT-542B
y . c c . . . BusTek BT-742A (EISA) (mPort specifies Revision F)
c . . . . . Compag SCSI Option Adapter and Compression Adapter
c . . . . . . Corollary SCSI-CPU
. . . c c . . DPT PM2102 caching controller (MFM emulation)
c . . c . . . DPT PM2102 caching SCSI controller in SCSI mode
. . c . . . . Everex EV8118/8110
c . c . . . . Future Domain 1660, 1680, 885, 860
y . . . . . . IBM HardFile (their SCSI host adapter for MicroChannel)
. . . c . . . Mylex DCE (EISA)
c . . . . . . Olivetti ESC-1 (EISA)
. . . . c . . PSI caching controller
c . . . . . . Storage Plus SCSI-AT "Sumo"
. . . c . . . Ultrastor 32k 12u
c . y c c c . . Western Digital WD7000
c . y . . . . . Western Digital WD7000-EX (EISA version of WD7000)
Notes:
* UHC started shipping a native-mode 1740/1742 driver in mid-April. It
requires a full SCSI-2 tape drive.
* The BusTek 542 is a clone of the Adaptec 1542. At least one respondent
thinks it works better and faster with the Adaptec drivers than the
Adaptecs do! The BusTek 742 has more complicated antecedents; it's an
EISA clone of the 1542, not necessarily compatible with the 1742.
* There's a known bug in the Adaptec 1742 firmware that produces hangs
when it's used with certain SCSI tape drives, including the popular
Archive 2150S.
* Bill Austin <uunet!baustin!bill> writes: "the 1740 patches on ESIX [4.0.3a]
do work but only bring the speed up in enhanced mode by about 15% over
standard (643Kb/s vs 535Kb/s) in writing, although the *read* speed
has nearly tripled (2,833 Kb/s) (this is using "iozone 16"). This may give
some idea of what improvement to expect from native-mode 1740 operation.
* Wolfgang Denk <w...@pcsbst.pcs.com> reports that SCO ODT 2.0 running an Adaptec
1542 cannot work with the following Hewlett-Packard drives:
HP 97536 SL
HP 97536 S
HP 97544
A source at SCO says "This problem is known to us. In some
not-yet-clearly-understood fashion, these HP drives interact badly with
our implementation of scatter/gather disk transfer ordering. There are
two different workarounds: you can turn off scatter/gather in the SCSI
disk driver, or you can get updated drive control board ROMs from HP."
S C D E M P U B X Network cards
-------------------------------------------------------------------------------
c . . . . . c y 3COM EtherLink I 3C501 and 3C502
c . c y c . c c 3COM EtherLink II 3C503
c . . . . . c . 3COM EtherLink 16 (3C507)
c . . . . . . . 3Com 3C523 & 523B EtherLink/MC
c . . . . . . . 3Com 3C523 EtherLink/MC TP
. . . c . . . . Everex EV-2015, EV-2016, EV-2026, EV-2027
c . . . . . . . HP 27245A EtherTwist Adapter Card/8 ISA TP
c . . . . . . . HP 27247A EtherTwist Adapter Card/16 ISA TP
c . . . . . . . HP 27250A ThinLAN Adapter Card/8 ISA BNC
c . . . . . . . HP 27248A EtherTwist EISA Adapter Card/32
c . . . . . . . IBM Token-Ring Network Adapter
c . . . . . . . IBM Token-Ring Network Adapter II (short and long card)
c . . . . . . . IBM Token-Ring Network Adapter 4/16
c . . . . . . . IBM Token-Ring Network Adapter/A
c . c . . . . . IBM Token-Ring Network 16/4 Adapter/A
c . . . . . . . Microdyne (Excelan) EXOS 205, 205T, 205T/16
c . . . . . . . Racal Datacomm NI6510 ISA and ES3210 EISA
. . y c c c . c Intel PC-586 aka iMX-LAN/586
. . . . . . c . Novell NE1000
. . . . . . c . Novell NE2000
y c y y c c c c c SMC & Western Digital 8003 and 8013 and variations
. . y . . . . . WD TokenRing card
Notes:
* SCO support of SMC EtherCards and the 3C507 requires a patch available
from their BBS.
* Dick Dunn <r...@raven.eklektix.com> opines "Somewhere along here, somebody
needs to note that the 3C501 is a miserable-misbegotten-son-of-a-lame-she-
camel-and-a-desperate-jackal Ethernet card, at least in UNIXland. It has
serious problems in any serious multi-user system because of various
hardware idiosyncrasies which are on the order of can't-walk-down-the-
street-and-chew-gum." Do tell, Dick!
S C D E M P U B X Tape drives
-------------------------------------------------------------------------------
c y y c y . c . Archive 2150S or Viper 150 21247 (SCSI, QIC-150)
c . c c . . c Archive Viper VP150E
c . . c c . . Archive Viper 60 21116
c . . c c . . Archive Viper 150 25099
c . . c c . . Archive Viper 2525 25462.
y . . c . . . Archive 60 - 525MB (QIC-02 and SCI)
c . . c . . . Archive 4mm 4520 DAT
c . . c c . . Archive Python models 25501-003, -005 and -008 (SCSI)
c . . . . . . Archive Python DDS 4520NT and 4521NT DAT drives
c c * c c . c Archive XL (5580 & friends)
. . . c c . . Archive 3800
. . . . c . . AT&T KS22762 and KS23495 (SCSI)
c . . . . . . Bell Technologies XTC-60
. . c . . . . Caliper CP150
c . . . . . . Cipher CP-60B, CP-125B
. . c . . . . Cipher ST150S-II
c . . . c . . Cipher ST150S2/90 (SCSI)
n . . c . . . CMS Jumbo - 60MB QIC-40
n . . . c . . Emulex MT02/S1 +CCS INQ (SCSI)
. . c c . . . Everex Excel Stream 60, 125, 150
. . c c . . . Everex5525ES (SCSI)
. . c c . y . Everex EV-811, EV-831, EV-833
c . . c c c . Exabyte EXB-8200 (SCSI)
c . . . . c . Exabyte EXB-8500 (SCSI)
c . . . . . . HP 35450A (SCSI)
. . . . c . . HP 88780 (SCSI)
. . . . c . . HPCIPHER M990 (SCSI)
. . . . c . . NCR H6210-STD1-01-46C632 (SCSI)
c . . . . . . Mountain 8mm Cartridge
y . . . n . . Mountain FileSafe 150MB (QIC-02)
c . . . . . . Mountain FileSafe 60-300MB (QIC-02)
c . y . . . . . Sankyo 525ES (SCSI)
. . . . c . . Sony SDT-1000 (SCSI)
. . . c . . . Tallgrass 150 - 525MB SCSI
c . . . . . . Tandberg DQIC (SCSI)
. . . . . . c TUV DAT
c . y . c . . . Wangtek 150SE (SCSI)
c . c c y . . Wangtek 5150ES (SCSI)
c . . c . . . Wangtek 60 - 525MB (QIC 02 and SCSI)
c . . c . . . Wangtek 6130 - HS 4mm DAT.
c . . y c . . Wangtek 5125ES ES41, 5150ES ES41, 5150ES FA0 (SCSI)
c . . c c c . Wangtek 5150ES SCSI-3 (SCSI)
c . . c . c . WangTek 5150PK QIC-02 (QIC-150)
c . y . . . . . Wangtek 5525 (SCSI)
c . . c c . . Wangtek 6130-F (SCSI)
c . . c c . . Wangtek KS23417, KS23465, KS24569 (SCSI)
Notes:
* All SVr4s inherit USL support for QIC-02, QIC-36 1/4", or SCSI tape
interfaces, using QIC-24 (9-track, 60MB), QIC-120 (15-track, 125MB) or
QIC-150 (18-track, 150MB) formats.
* A user says of Dell: it appears that anything using Wangtek QIC02/QIC36
controllers works; this should include the Wangtek 525MB, Cipher ST150S2,
and Archive 2150S drives.
* UHC specifies the following tape controller/drive combinations: Wangtek
PC-36 + Wangtek 5099-EN, Everex 811 + Wangtek 5150-EN, Bell Tech + Wangtek
5150-EN, Archive SC499-R + Archive External FT-60, Archive VP402 + Archive
Viper 2150L, Everex 811 + Archive Viper 2150L, Bell Tech + Archive Viper
2150L, Archive VP402 + Archive Viper 2150L.
* UHC claims that Any floppy tape supporting the QIC-107 physical and QIC-117
logical interface specs and QIC-80 or QIC-40 recording formats should work.
This is probably true of other vendors as well.
* BSDI says it supports almost any Wangtek 1/4" standard 3M streamer with a
QIC-02 or QIC-36 interface. However, they admit that the Archive SC402
QIC-02 controller will not work. BSDI says it will support almost any SCSI
tape unit, as well.
* Floppy tapes don't work on Dell; USL provides the support, but it collides
with Dell's code for auto-detecting the density of a diskette.
* SCO's tape compatibility table lists drive/controller pairs; not all drives
listed have been included here. They allege that any QIC-02 drive should
work. Unofficial sources inside SCO claim any SCSI drive ought to work.
* A source at SCO says the CMS Jumbo is neither compatible with QIC40/QIC80
nor Irwin "standards", vendor supplies their own driver which SCO does not
support. He also said "CMS is in general fairly UNIX-hostile; don't buy
their stuff if you have a choice." On the other hand, Jerry Rocteur <jerry
@lncc.com> praises their hardware and says he found them quite helpful and
knowledgeable. Your editor has no experience on which to base an opinion.
* The Emulex MT02 is a QIC02 bridge controller for the SCSI bus -- lets you
take an old QIC02 drive and run it on a SCSI bus. It is said to use a
very old version of the SCSI spec; caveat emptor.
* John Plate <pl...@infotek.dk> writes: "According to a fax from the Archive
manufacturer Maynard, [the XL 5580 drive only works with ESIX 4.0.3] if the
tape drive is "drive" two! Which is the same as disabling the second floppy
drive and then set a jumper on the tape drive."
S C D E M P U B X Non-Winchester mass storage
-------------------------------------------------------------------------------
. c . . . Bernoulli 90MB exchangeable SCSI
. . c . Hitachi, Toshiba (models not specified)
. . . c Maxtor RXT-800HS
. c c . Storage Dimensions XSE1-1000S1 optical disk
. y c . SyQuest cartridge media
. c . . . Tandata
. c c c Toshiba TXM-3201A1 CD-ROM
. c y c c Toshiba TXM-3301B CD-ROM
. . c c Toshiba WM-C050
. c c c Toshiba WM-D070 WORM drive
VII. FREEWARE ACCESS FOR SVR4 SYSTEMS.
US4BINR is an archive dedicated to binaries (executable compiled program)
for UNIX System V Release 4 (SVR4) on 386/486 PC computer.
Our goal is to provide easy access to precompiled programs. Those
programs are (hopefully):
Up to date.
Documented.
Useful or fun.
Uploads annoucement are made in comp.unix.sysv386 and comp.unix.sys5.r4.
US4BINR carries PD, Freeware, shareware, games, etc... US4BINR is a non profit
organisation.
To get more info, email the following message to requ...@us4binr.uucp
or request%us4binr.u...@uunet.uu.net
reply Put_your_email_address_here
help
quit
VIII. FREE ADVICE TO VENDORS:
As a potential customer for one of the SVr4 ports, it's to my advantage to
have everybody in this market competing against one another as hard as
possible. Accordingly, some free advice to vendors, which I'm broadcasting to
all of them and the public so as to put just that much more pressure on each
vendor. :-)
SCO:
You have a serious image problem with many hackers which you've exacerbated
recently by falling behind the SVr4 leading edge and then engaging in what
certainly appears to be an attempt to sucker careless buyers with deceptive
product naming. But the reaction to this wouldn't be nearly so vehement if
it didn't come on top of years of discontent with more technical choices.
There's too much stuff in the SCO kernel and admin tools that's different from
USL and *not better*; too much stuff that raises weird little compatibility
problems that shouldn't be there. Verbum sap.
This different-but-not-better problem is perfectly reflected by the one
thing about the otherwise-excellent SCO documentation that sucks moldy moose
droppings; the rearrangement and renaming of the reference manual sections.
Your technical writers entertain a fond delusion that this helps nontechnical
users, but all it really does is confuse and frustrate techies with experience
on other UNIXes. Lose it.
Everybody but SCO:
SCO's documentation set is to die for (except in the one respect noted
above), and they add a lot of value over the base UNIX with things like ODT DOS
and CodeView. Only Dell comes even close to matching SCO in the nifty add-ons
department, and even they have a lot of room for improvement. If you want to
outcompete SCO, you have to be *better*; this means (at minimum) supporting a
windowing debugger and ISAM libraries and DOS support that goes beyond 2.0.
Consensys:
Fix the Powerports bugs everyone is reporting. They're doing you real
damage. Nobody expects real support from an outfit selling at $1000 below
market average, but you've *got* to make your own hardware work right or look
like idiots.
Beyond this, I think you have a serious attitude problem. So far, you're
the only outfit out of nine to refuse to divulge information for the
comparison tables. While you have a perfect right to do so, it smells bad ---
as though you think you have weaknesses to hide. I tried to discuss this with
your VP of sales (Gary Anderson) and got back very little but evasions,
suit-speak, defensiveness, and attempts to divert me from the issues (and I
don't mind admitting that the conversation made me pretty angry and didn't end
very pleasantly). This man's behavior is all too consistent with reports of
Consensys's dismissive behavior towards customers and continued refusal to
acknowledge technical problems.
In this corner of the industry we have a tradition of collegiality, mutual
trust, informality, and candor. If you plan to be here for the long haul, you
need to learn how to work with that rather than fighting it. Behaving like IBM
will only get you hammered.
Consensys and Esix:
Get a real support address. Bang-path accessibility doesn't impress anyone
any more --- in fact, it looks faintly quaint. You guys ought to be
supp...@everex.com and supp...@consensys.com to follow the simple and logical
convention SCO and Dell and Microport and UHC have established.
Dell:
Don't get fat and lazy. You've got the lead in the SVr4 market at the
moment and you've got the money and resources to keep it, *if* you use them.
If you staff up your UNIX support operation so customers don't get pissed off
by infinite hold, *and* keep your prices the lowest in the upper tier, no one
will be able to touch you. Don't let Microport et al. get ahead of you in
releases and new technology, and try to reverse that creeping corporatitis (the
no-comment-on-unreleased-products policy is a bad sign).
Everybody but Dell:
Offer all the free software Dell does --- and *more*. All it will cost you
is the media, right? Even if you have to plaster CONTRIBUTED SOFTWARE, NOT
SUPPORTED on it, include perl, elm, bison, gcc, emacs, gdb, mush, patch,
compress, etc on your distribution tapes. Heck, include some *games* (SCO
includes games with UNIX but not the full ODT product; and makes some games
available for download on their BBS).
Nethack, empire, zork, stuff like that. Your engineers use and
play with all this in-house anyhow, yes? And you're selling to guys
just like your engineers. They'll love you for it. Trust me.
Set up a `sales' address to take product queries if you don't already have
one.
Everybody but Dell and SCO:
A Dell person warns that the kinds of tweaks to the source made by porting
houses can break X/Open (XPG3) conformance. Dell and SCO test every build with
VSX (the X/Open-approved XPG3 test suite) and Dell reports that it often finds
places where seemingly innocuous bug fixes cause XPG3 violations. Other UNIX
vendors would be well advised to do likewise.
Set up an 800 number for tech support. Support customers hate spending time
on hold, and they hate it like poison when they have to *pay* for the hold
time. The more overloaded your support staff is, the more important this
gets. Verbum sap.
Esix:
You're *boring*. You seem to make a decent product, but there's nothing
I've seen about ESIX that'd make me say "I might want to buy ESIX because...".
Position yourselves; pick something like price or support quality or
reliability or add-on features and push it hard. Warning: if you decide to
push support, *hire more engineers*. Your rep for following up on support
problems is bad enough that your "unlimited free support" ain't much of a draw;
especially now that your two best support guys have quit.
Esix, MST, UHC:
Get 800 numbers for product info, too.
MST:
Set up a supp...@mst.com alias to your cs address, see above. What would
that take, a whole five minutes? :-)
If you don't start planning for 4.0.4 now, you'll get left behind this
spring and early summer whan all the other vendors move to it.
On present trends, your software prices are cheap enough; you'd probably
get more sales mileage out of pulling down the hardware prices for your
pre-configured systems.
Everybody but MST and Microport:
Set up a `sales' alias to your info and orders email address. A universal
convention for this means just one less detail prospective customers need to
remember.
Microport:
Your complete system is way overpriced relative to what other vendors in the
top tier are selling. If I were a corporate customer, there is no *way* I
could justify spending the $1K or $2K premium over Dell's price --- not when
Dell has the rep it does for quality and features. You aren't offering
anything but a crippled copy of JSB Multiview to justify that premium and
that ain't enough.
There's some evidence that you've got a technical lead on the competition.
Push it; push it *hard*. You're first off the blocks with 4.0.4; keep that up,
be first out with a stable 4.0.5. Market yourselves as the leading-edge
outfit, court the hard-core wizards as their natural ally, detail somebody
who's fluent in English as well as C to listen and speak for you on USENET, and
keep the promises you make there.
UHC:
You've decided to push support; that's good, but follow through by getting
that 800 number. Don't lose those small-company virtues of candor and
flexibility, trade on them. Your policy of having all techs clear up to the
product manager take turns on the support lines is a damned good idea, stick
with it. And I'm sufficiently impressed with what I've heard from your guys
that I think you might be able to fight Microport for the friend-to-wizards
mantle, too. Maybe you should try.
Everybody except BSDI:
BSD/386 includes *sources*. For *everything*. Be afraid; be very afraid.
In effect, this recruits hundreds of eager hackers as uncompensated development
and support engineers for BSDI. Don't fool yourselves that the results are
necessarily going to be unfocused, amateur-quality and safe to ignore --- it
sure didn't work that way for gcc or Emacs. The rest of you will have to work
that much harder and smarter to stay ahead of their game.
BSDI:
Don't you get complacent either. The 386BSD distribution is breathing
down *your* neck...
The most effective things you can do to to seriously compete with SVr4
vendors are: a) emphasize standards conformance --- POSIX, FIPS, XPG3, etc.,
and b) follow through on your support promises. Just another flaky BSDoid
system isn't really very interesting except to hobbyists, even with sources ---
but if it were proven a reliable cross-development platform it could capture
a lot of hearts and minds among commercial software designers.
Everybody:
Do something about your product names! Even the cases that don't appear
to be deliberate deception are very confusing to the customer. If you're
releasing an enhanced 4.0.3 or 4.0.4 that's what you ought to *call* it. I
recommend:
Consensys UNIX Version 1.2 --> Consensys UNIX 4.0.3 revision 1.2
Dell UNIX Issue 2.1 --> Dell UNIX 4.0.3 revision 2.1
ESIX System V Release 4.0.4 --> Esix UNIX 4.0.4 revision 4
MST SVr4 UNIX --> MST UNIX 4.0.3
Microport System V/4 version 4 --> Microport UNIX 4.0.4
UHC Version 3.6 --> UHC UNIX 4.0.3 revision 6
The fact is, all these idiosyncratic version-numbering systems do you no
good and considerable harm. At worst, they make it look like you're trying to
pull a scam by deceiving people about the level of the base technology. At
best, they parade your internal revision number (which conveys no useful
information unless one is an existing customer considering an upgrade already)
and obscure the really important information. Do your product differentiation
elsewhere, in substance rather than nomenclature; it's not useful here.
You're *all* badly understaffed in support engineering, and it shows. Boy
does it show --- in poor followup, long hold times, and user gripes. The first
outfit to invest enough to offer really first-class quick-response support is
going to eat everyone else's lunch. Wouldn't you like to be it?
IX. ACKNOWLEDGEMENTS AND ENVOI
Some of the material in this posting was originally assembled by Jason
Levitt <ja...@cs.utexas.edu> of "Open Systems Today". Grateful acknowledgement
is made to him for permission to re-distribute and update this information.
Many netters sent me email contributing technical information, feedback,
and comment. Thanks to all. It's in combinations of individual mission and
collective cooperation like this one that the net really shines, and I'm
grateful to everybody who's worked with me to improve the signal/noise ratio.
The level of cooperation I've experienced from vendors' program managers,
techies and marketing people since the first issue has generally been
outstanding. Particular high marks go to Jeremy Chatfield at Dell, Kristen
Axline at Microport, John Prothro and Sam Nataros at UHC and Bela Lubkin at
SCO, with very honorable mentions to Jeff Ellis at Esix and Rob Kolstad at
BSDI. By cooperating intelligently with this FAQ, they've done a great job of
serving the market and representing their corporate interests.
One dishonorable mention goes to Gary Anderson, V.P of sales at Consensys
and the only person I've encountered who's behaved like the classic stereotype
of the slippery, stonewalling marketroid. An impression of this kind is
exactly what Consensys needs to solve their credibility problems...NOT!
So far, I've found that the technical merit of each of these eight products
(insofar as I have data to judge; I haven't actually used any of them yet)
seems to correlate pretty well with the degree of cooperation I've received. I
wasn't explicitly expecting this result, but I'm not surprised by it either.
--
Send your feedback to: Eric Raymond = e...@snark.thyrsus.com