Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

SUMMARY: PCI motherboards for OS/2

3 views
Skip to first unread message

Pat Duffy

unread,
Jul 24, 1995, 3:00:00 AM7/24/95
to
Document: PCI Motherboards for OS/2 list
Maintainer: Patrick Duffy, du...@theory.chem.ubc.ca
Last Revision Date: July 23, 1995
Archived at: ftp.netcom.com, in directory /pub/ab/abe/
Web pages: http://warp.eecs.berkeley.edu/os2/workbench/work.htm

This is the PCI motherboards for OS/2 list. For detailed descriptions
of PCI chipset problems and workarounds, please refer to the PCI
chipsets list. As usual, '*' indicates the recommended motherboards,
and '**' indicates that I would choose this motherboard myself when
picking a PCI motherboard for my own PCI system.

Also as usual, please feel free to correct any continuing misconceptions
I might have, to insist that I should recommend a motherboard I don't,
or to add a new motherboard/bit of information to what's here already.
Your contributions and/or corrections are always welcome and certainly
appreciated. Please, when reporting problems or successes, try to be as
detailed as possible in your hardware descriptions. BIOS rev. numbers
are especially useful.

Useful Numbers:
---------------
AIR: (408) 428-0800
Asus: (408) 956-9077 (tech. support)
ftp.asustek.asus.com.tw (ftp site)
http://asustek.asus.com.tw/Products/catalog.html (WWW site)
gopher.asus.com.tw (gopher site)
t...@asus.com.tw (tech. support E-Mail)
alt.comp.periphs.mainboard.asus (internet newsgroup)
Award: (415) 968-4433 (Voice)
(415) 968-0274 (FAX)
(415) 968-0249 (BBS - 8,N,1)
1:143/2...@fidonet.org
EliteGroup: (510) 226-0434 (faxback)
Gigabyte: (818) 854-9338 (tech. support)
Intel: (800) 628-8686 (tech. support)
(916) 356-3600 (BBS -- N,8,1)
+44-793-496340 (U.K. BBS -- N,8,1)
(800) 628-2283 (faxback -- order catalogue 7 for a list
of PCI-related documents)
J-Bond: (408) 946-9622
Lexar: (408) 748-9199 (voice)
(408) 748-1040 (FAX)
Micronics (510) 651-2300 (Office -- Faxback after hours)
(510) 651-6985 (BBS)
Washburn (AMI): (800) 836-9026 / (716) 248-3627 (General inquiries)
(800) 836-8027 (Motherboard Hotline)
(800) 836-8028 (faxback and information about specials)
(716) 383-6086 (tech. support)
(404) 246-8600 (tech. support)
(716) 381-7549 (FAX)
(404) 246-8780 (BBS - V.34)
(404) 246-8781 (BBS - V.34)
(404) 246-8782 (BBS - V.32 or HST)
(404) 246-8783 (BBS - V.32)


Manufacturer Model Comments
------------------------------------------------------
AIR 486VP This is AIR's PCI/VL/ISA '486
board. It uses the Contaq chipset
and the Award (4.50g) flash BIOS. It will take all the different CPUs,
up to the DX4/100s and Pentium Overdrive processors. It comes with 256
kB of 15 ns cache. The board has been reported to work properly under
OS/2 2.1/2.11/Warp. This board has, apparently, recently been
discontinued.

Setup tips: None so far.
Possible problems: None so far.

AIR 54CEP This is AIR's PCI/EISA 90/100 MHz
board. It uses the AMI BIOS (flash
optional), and the Mercury chipset (I'm not sure why it uses the Mercury
and not the Neptune chipset). It will accommodate 5 PCI cards and 4 EISA
cards, all of which support busmastering. The board will take up to 128
MB of 72-pin SIMMS, and either 256 or 512 kB of L2 cache. The board has
all the green features. There is a fast or fast/wide PCI SCSI port
(both connections are present, apprently) built in which is based on the
Adaptec 7870 chip. 2.88 MB floppy drives (up to two) are supported, as
are 2 16550 serial ports, one mouse port, and one parallel port. I've
had one report of success with this board using DOS/Win and Netware 3.12
(not surprising), as well as Warp.

Setup tips: None so far.
Possible problems: Quantum 1 GB drives do not communicate at full speed
with the Adaptec 7870 SCSI chip. It is necessary
(if you're using this drive with that chip) to set
communication to 8 MB/sec.

ALI PCI P5-60/66 This motherboard has 4 PCI and 4
ISA slots, one of each of which is
shared, meaning you can run it as 4 PCI/4 ISA or 3 PCI/5 ISA. It does
not have built-in support for the NCR 53c810 chip, so you'll need a
53c825-based controller if you want NCR. The motherboard uses the ALI
M1449 chip to support ISA/PCI and Green standards, and the ALI M1451
chip to provide Host/PCI bridge. The 66 MHz version has been reported
to work without problems with both OS/2 2.1 and OS/2 Warp, though not
with Warp Beta 2 (it did work with Warp Beta 1). The 60 MHz board has
been reported to work with Linux.

ALI J624 This is ALI's 90 MHz board. It has
all the green features, and uses
the AMI graphical BIOS. The board has been reported to not work 100%
reliably with Warp.

AMI Super Voyager This is AMI's 486 board. It uses
PCI II the SIS chipset and will
accommodate up to a Pentium
Overdrive processor. It is, apparently, plug-and-play 1.0A-compliant.
The board will accommodate up to 128 MB of 72-pin SIMMs on board, and
comes with 128 kB (upgradable to 256 kB) of cache. The flash BIOS (by
AMI, of course) uses the WinBIOS interface, and supports IDE,
auto-configuring of PCI slots, and all the green features. The PCI
slots (there are three) are all busmaster-enabled and 2.0-compliant.
There are four ISA slots. Presumably there is one shared PCI/ISA slot,
though I've no indication of this. Floppy/serial/parallel support is
built in. I've had no reports of success with this board under OS/2 as
yet.

Setup tips: None so far.
Possible problems: None so far.

AMI ATLAS PCI This is AMI's 90/100 MHz Pentium
motherboard. It uses the SIS
chipset and will take up to 128 MB of RAM in 72-pin SIMMS and 512 kB of
cache (256 kB is standard). The BIOS (AMI) supports all the green
features and advanced IDE modes. There are four PCI 2.0-compliant
slots, all of which allow busmastering, and four ISA slots (one of each
of these slots is shared). (E)IDE/serial/parallel/mouse support are all
built-in as well. The BIOS will automatically configure the PCI bus
and is plug-n-play 1.0A-compliant. The board is reported to work well
with OS/2.

Setup tips: AMI has apparently released a BIOS for this
motherboard which is specific to S3-based cards.
Apparently the motherboard will not recognize cards
like the Stealth 64 Video VRAM unless this BIOS is
in place.
Possible problems: None so far.

Asus PCI/I-486** This board uses the Saturn rev. 4
SP3G chipset, and will accommodate both
the Intel and Cyrix CPUs, up to the
DX4s. In addition, the socket will also accommodate a P24T/P24D. It
also has all the 'green' features. It co-exists with the SP3, but since
it has the newer Saturn chipset I'm recommending it instead. (Make sure
to be very clear when specifying your motherboard that you want the SP3G
and not the SP3.) The board (currently at rev. 1.8) will accommodate up
to 128 MB of RAM (four sockets which must be filled in pairs) and 512 kB
of write-through L2 cache (256 kB is standard). This board has the NCR
53c810 SCSI controller on-board (with a standard internal 50-pin socket
for internal SCSI devices), as well as super multi-I/O
(IDE/serial/parallel) and BIOS support for 2.88 MB drives. There is
BIOS support for up to four IDE drives, though the board will only
accommodate two (on the ISA bus). (Internal IDE must be disabled and an
EIDE controller obtained if support for four EIDE drives is desired.)
The board has 4 ISA/3 PCI slots (one slot is shared between the PCI and
ISA bus and so effectively you have 5/2 or 4/3 slots) and a built-in
floppy controller. There is also a socket for a mouse (either a
header-style socket or a PS2 style) which takes up IRQ 12 if enabled.
I've had reports of success with OS/2 2.1, 2.11, Warp Beta II, and Warp
GA and this board. The AWARD Flash-BIOS on board is at revision 4.50G,
and the NCR .ADD file is dated 28/4/94. It should also be noted that
the jumpers in these motherboards must be set up carefully, and by
consulting with the manual which comes with the board. Apparently, Asus
has just released an update to the flash BIOS for this board. Everything
is reported to work, with the exception that OS/2 apparently does not
start from the boot manager. OS/2 still boots from floppy, and
downgrading to the 301 bios fixes the problem.

Setup tips: All 3 PCI slots on this board are fixed at
PCI INT A for level triggering assignment. IRQ to
PCI INT for each PCI slot is done in the BIOS. For
edge trigging, assignment of the IRQ is done with
the on-board jumper settings for the actual slot.
If you're running the board with an AMD DX4 (3x33)
CPU, it is necessary to set your jumpers as for a
non-SL enhanced DX4, except set J36 to 1&2 rather
than 2&3. The AMD DX4 will run in 2x mode (2x33) if
pin B13 is not grounded.

Possible problems: apparently the on-board SCSI-controller has problems
co-existing with OS/2 2.1 and a Quantum Prodrive 540S,
as synchronous communication must be disabled in order
for the system to boot. Other drives seem to work
well with the NCR chip (I have a Quantum Empire
1080S).

Asus PVI/486AP4* Asus may finally have it right with
their line of 486 PCI boards. This
particular board has the Aries (rev. 2) chipset and was reviewed very
favourably by C't. This board will take all the different 486 chips,
including the new DX4 (at 75 or 100 MHz), and has a spot for a Pentium
Overdrive P24T via a ZIF socket. It comes with 256 kB of cache
(write-back, L2), and will accommodate 128 MB (4x32 MB 72 pin SIMMs). It
has the latest Green features (Award BIOS, etc.), and has the NCR SCSI
BIOS built in (though there is no 53c810 chip itself). EIDE (PCI) is
built in as well. The board has 1 combination ISA/VL/PCI slot (only one
of the three slots may be used), 3 ISA and 3 PCI slots. I have had a
couple of reports of success with this board and OS/2 (2.1/2.11/Warp)
and NT. Revision 1.6 of the board still requires you to use the reset
button to reboot your machine if you have a SCSI controller installed.

Setup tips: One person has suggested that it would be better to
not use the VL slot in the board (to just disable
it). Upon doing this, the board is said to be very
stable. If you get an NCR SCSI card, put it in slot
1 to get the system to boot.
Possible problems: One person has reported that the chipset ID
procedure given in the chipset list does not work
for this board.

Asus PVI/486SP3* This motherboard uses the SIS
chipset and takes all the different
486 CPUs. Like the SIS Pentium chipset, the chipset used here will
allow many different external clock settings, so that DX2-80s and DX-40s
are well-supported. The board has 3 PCI slots, 3 16-bit ISA slots, and
a (shared) PCI/VL slot, and all the on-board integrated I/O (2 VL IDE
ports, 1 floppy port, 2 serial poarts, a mouse port, and one ECP/EPP
parallel port). The board uses the Award BIOS (which has the NCR SCSI
BIOS built in), though a Flash EPROM is apparently only optional. The
board will take up to 128 MB of RAM (in two sockets, if you can find a
64 MB SIMM), and will allow only one socket to be filled. The board is
reported to work very well under DOS, Windows (3.1), and Warp.

Setup tips: None so far.
Possible problems: The B2 revision of the SIS chipset apparently does
not support mode 3 IDE well. Running a DX4 CPU at
50 MHz instead of 25 is reported to improve
performance substantially.
Useful information:

This board comes in three slightly different varieties, corresponding to
different revisions of the SIS chipset used: B2, B3, and A4 The A4 does
not have onboard ide support. The B2 and B3 do; the B3 is the latter
revision of the two. The presence of the B2 chipset can be verified by
looking at the large chipset chip with "SIS" marked on it; It should
have the letters "NU" on the third line of text (counting the SIS in big
letters as a line). B2 apparently does not support mode 3 IDE well, and
B3 is supposed to fix this.


Asus PCI/E-P5MP3* This particular motherboard has
PCI/EISA been reported to work well, with
the one caution that older
motherboards had a bug in the serial I/O hardware. This board has
identical specifications to the motherboard below, with the exception
that this board uses EISA slots instead of ISA.

Setup tips: None so far.
Possible problems: None so far.

Asus MB-586A- This is Asus' bare-bones 60/66 MHz
PCI60C PCI board. It does not have any
I/O on board, but does have a ZIF
socket for a future upgrade. 256 kB of L2 write-through cache is
standard, with 512 kB an option. The board uses the Mercury chipset
(Intel) and the Award Flash-BIOS. It has 4 ISA and 3 PCI slots, and
will accommodate up to 6x32 MB 72-pin SIMMs.

Setup tips: None so far.
Possible problems: None so far.

Asus PCI/I-P54NP4 This is an ISA-based board which
takes up to two 133 MHz CPUs with
the latest BIOS update from Asus' web site. It has all the standard I/O
built in (serial/parallel/IDE). It has 256 kB of on-board cache,
upgradable to 512 kB. It will take four 72-pin SIMMs (with parity), and
has a flash BIOS. The board uses the Neptune chipset. I have had a few
reports of success with various revisions of this board and OS/2
2.11/Warp. The board may now work with OS/2 SMP with the latest BIOS.

Setup tips: The board has many jumpers and, apparently, nearly no
documentation to tell you how to set them. (Two pages
of photocopied jumper settings, I'm told, are all that
you get.) One other person has written to say that
you get a very good manual with the board, so it
could be that the documentation you get depends
largely on where you buy the board. Get the latest
BIOS to ensure correct operation and support for 133
MHz CPUs and OS/2 SMP.
Possible problems: The board will not work with two CPUs and OS/2 SMP.

Asus PCI/E-P54NP4 This is Asus' dual-processor board
which uses the Neptune chipset. It
has been tested under a variety of operating systems and, until
recently, the only one under which it did not work was OS/2 SMP. A BIOS
update (from Asus' web site) will apparently now make it work with OS/2
SMP and (up to two) 133 MHz CPUs. I've no technical specifications for
the board, other than that it is a combination PCI/EISA board, and uses
the Neptune chipset.

Setup tips: Get the latest BIOS for the board for trouble-free
operation under OS/2 SMP and support for 133 MHz
Pentium CPUs.
Possible problems: None so far.

Asus PCI/I-P54SP4* This board can take a 75/90/100 MHz
Intel Pentium CPU, accommodates
256/512/1 MB of L2 cache and uses the SIS PCI chipset. The AWARD BIOS
(4.50g -- flash upgradable) is used, which has the NCR SCSI BIOS built
into it. There are four 72-pin SIMM sockets for RAM, so the board can
take up to 128 MB. The board has two 32-bit PCI IDE ports (via the CMD
640B chipset -- up to four drives may be attached), 1 floppy port (2.88
MB support built-in), two 16550 serial ports, and one ECP/EPP parallel
port. It can be run with 3 PCI/4 ISA or 4 PCI/3 ISA slots. A beta
release of this board has apparently run OS/2 2.1 and Win/NT with no
problems, and revision 1.4 (the current version) apparently runs Warp
(and Warp full-pack) without problems. All PCI slots in this board
allow busmastering. People have reported problems with this board not
booting OS/2 with various PCI SCSI controllers; these problems were
solved by disabling the green functions in the BIOS. The same problems
are not reported by another person with a 2940W, however. I have also
had reports of success using an IDE drive in combination with the new
ATAPI IDE CD-ROM, on which Warp full-pack was reported to install
without problems (after modification of the install disks to point to
the new ATAPI driver, of course).

Setup tips: The PCI bus speed on the board apparently may be set
in the BIOS setup as either CPU CLK/2 or CPU CLK/1.5
(to support the 75 MHz chips or, alternatively, to
overclock the PCI bus). I don't know what effect
running the PCI bus at higher-than-spec rates would
have, though. The BIOS apparently also has many
settings for PCI bursts and wait-states. Setting
all of these to the fastest allowable values seems
to work. The parallel port is assigned to LPT2: by
default in the BIOS. The pin-out for the on-board
mouse connector is:

------------------------------
| Gnd Data N/C +5 |
| |
| Clk -- -- N/C |
------------------------------

On a PS/2 mouse, these correspond to the following
signals:

Pin # Signal
----- ------
1 Data
2 N/C
3 Gnd
4 +5
5 Clk
6 N/C

The "Computer Stop" (206) 644-5400 apparently sells
pre-made mouse cables.

Possible problems: The green functions in the BIOS are apparently
incompatible with many SCSI controllers; it may be
necessary to disable them in order to get Warp to
boot at any resolution beyond VGA. See the PCI IDE
list for a description of problems with the CMD PCI
IDE chipset. Some people have been having problems
with compressed files being corrupted as they are
uncompressed from floppy (a Diamond Stealth VRAM was
in use on both systems, though this may just be
coincidence).

Asus PCI-I/P54TP4** These are the boards by Asus
PCI-I/P55TP4* which use the Intel Triton
PCI-I/P55TP4XE** chipset, with all that that
entails. The boards themselves
(note that the P55TP4 has been discontinued) take a 75/90/100/120/133
MHz Pentium, and can be run with either 4 PCI/3 ISA or 3 PCI/4 ISA slots
(all of the PCI slots allow busmastering). Apparently one of the four
PCI slots is a proprietary Asus Mediabus, which means that any cards
designed to be run in it may have multiple functions (the combination
sound/video card detailed in the video cards list would be one of
these). The board will take from 8 to 128 MB of DRAM or EDO RAM, and has
256 kB of SRAM cache (upgradable to 512 kB), which may be either
synchronous or asynchronous. The synchronous cache is activated and the
asynchronous deactivated automatically when SRAM is plugged into the
socket provided for the synchronous cache. It has the SMC super I/O
controller on board, and PCI EIDE via the Triton chipset, which
apparently supports mode 4 data transfers and DMA mode 2. The board uses
the Award BIOS (which has the NCR BIOS built in). There is also a mouse
port, but no turbo switch connector on the board. The board was reviewed
in the 4/95 issue of the German computer magazine C't. The review was
very favourable towards the new Triton chipset and the ASUS board they
tested. A board equipped with 256k Burst-SRAM and EDO-RAM achieved
transfer rates of 65 MB/sec to 2nd level cache, 39 MByte/s on a direct
memory access, 53 MByte/s on a write operation (STOSD), and 54 MByte/s
on a memory to PCI transfer. Application benchmarks were run under
Windows, OS/2 Warp, and Windows NT 3.5. They didn't mention any
incompatiblities with PCI-components. The board failed when running the
PCI BIOS check, which was attributed to the BIOS. (The PCI/I-54NP4 board
which they used as a comparison failed that test as well). I have had
several reports of success from people running these boards.

Setup Tips: The BIOS settings for the board are apparently
complex and the documentation minimal. Some time
may be required to get it set up just right. To
that end, make sure you have the latest BIOS to
ensure trouble-free operation.
Possible Problems: One person has reported problems with this board and
a RAM chip labelled Ti -60 TMS417400DJ VBP 440230,
wherein NT/3.5 would neither install nor run on the
system. The board does a PCI bus reset after the
SCSI BIOS scans its bus, which causes problems for
the QLogic SCSI controllers. Make sure that if you
are using this combination of hardware that you get
the fix for the problem from QLogic (the ISP1020
firmware level should be 1.27 or greater).

Useful information:

These Web sites contain some technical information on the Asus boards:

http://www.infinet.com/~venkat/ (finger ven...@infinet.com)
http://www.tdl.com/~netex/ (finger ne...@tdl.com)

DTK QUIN-35 This board uses the Award BIOS
(flash optional) and the SIS
85C501, 85C502, and 85C503 chipset. In addition, motherboard
specifications list the WinBond W83769F, W83787F, and W83768F chips,
though I'm not sure what they do. The board is PCI 2.0-compliant. The
board will accept up to 1 MB of standard cache, and 128 MB of
conventional DRAM in four 72-pin sockets. The board has 3 PCI, 3 ISA,
and one shared PCI/ISA slot, as well as the (nearly standard) 2 serial
(16550), 1 EPP/ECP parallel, and one game port. Support is provided for
up to a 1.44 MB floppy drive, and 2 EIDE drives. I've had no reports of
success or failure with this board and OS/2.

Setup tips: None so far.
Possible problems: None so far.

EliteGroup SA486P AIO-U Uses the Saturn chipset and has both
(STD) IDE and NCR SCSI onboard. Current crop
now has revision 4 of the Saturn
chipset. OS/2 2.1 has apparently been installed with all caches on and
runs (using the NCR controller) with no problems according to reports.
Apparently too rev. 0.4 of the board hangs if the cache is set to
"write-back". This board uses the SMC 37C665 I/O controller, so make
sure yours has 'GT' at the end of the model number, to fix problems with
system crashes when using comm. programs.

Setup tips: The board has many jumpers and, apparently, nearly no
documentation to tell you how to set them. (Two pages
of photocopied jumper settings, I'm told, are all that
you get.) It may be necessary to set the on-board
NCR SCSI controller to IRQ 15 to get it to work.
Possible problems: Apparently the BIOS that ships with the board (burned
in, not flash) will not allow the user to change the
settings for 'Host-to-PCI'-Posting,
'Host-to-Memory'-Posting, or 'PCI-to-memory'-posting,
altthough this is possible via CTPCI.EXE, a small
program which is avaiblable from the German magazine
C't. (Flash BiOSes are available as an option.)

EliteGroup UM8810P AIO This board takes all the Intel and
Cyrix CPUs, including the Pentium
Overdrive series. It takes up to 64 MB of RAM (in either 4x16
single-sided or 2x32 double-sided) and takes up to 512 kB of L2 cache.
There are 3 PCI slots (all of which permit busmastering, one of which is
shared) and 4 ISA slots (one of which must also be shared, presumably).
It uses the CMD chipset for PCI IDE, and the SMC chip for built-in
serial/parallel I/O. Support is built-in for up to 2.88 MB drives. The
board uses the Phoenix BIOS and supports all the green features. The
board uses the UMC 888X chipset for PCI support. I've had a report of
success with Warp and this board.

Setup tips: None so far.
Possible problems: The BIOS allows you to run the PCI bus at either
half or all of the external CPU clock. This might
cause problems for DX40s, where you'll either be
able to run it at 20 or 40 MHz. The ISA bridge
(using these CPUs) will also not allow you to set
the bus speed to close to 8 MHz.

Gigabyte GA-486IM This board takes all Intel and
Cyrix CPUs and uses the UMC 888X
chipset. It uses 256 kB of L2 cache and the Award flash BIOS (version
4.50B, dated 5 Dec. '94). Presumably the board can take up to 128 MB of
RAM. This particular board does not work with an NCR SCSI controller
made by Intel or Asus, although it does work with an external EIDE
controller.

Setup tips: None so far.
Possible problems: Many. The board will not boot OS/2 at all with an
NCR controller made by Intel or Asus with a Cardex
Challenger in a PCI slot. A switch to an S3/864
card will allow the system to boot, but the serial
ports go undetected, or at best work poorly.

Gigabyte GA-486IS This board uses the Saturn I rev. 2
chipset and has the NCR on-board
SCSI chip. It will accept up to four 72 pin SIMMs (parity or no
parity), and has 4 PCI and 4 ISA slots. The clock speed is switchable
between 25 and 33 MHz, and the board supports only 5 V CPUs. There is a
ZIF socket on board which will accept a 486 SX, DX, DX2, or P24T chip.
The board has 256 kB of L2 cache. OS/2 reportedly runs well (with no
problems) after upgrading the original BIOS, which had compatibility
problems.

Setup tips: Disable the external cache on this board for
reliable operation under OS/2, especially if you use
a PCI-based SCSI controller.
Potential problems: See the PCI chipset list for a description of the
problems with the Saturn I rev. 2 chipset.

Gigabyte GA-586AL/S This is Gigabyte's 60/66 MHz
motherboard. It uses the Award
BIOS and ALI chipset. I don't have any other specifics on the board or
chipset, but apparently the board does _not_ work well with OS/2. The
system will apparently install well, but on reboot cannot locate the
desktop. To be avoided.

Gigabyte GA-586AP This board will take either a 75,
90, or 100 MHz Pentium. It uses
the ALI chipset and the Award Flash BIOS. The board has 2 dedicated PCI
slots, 3 dedicated ISA slots, and one shared ISA/PCI slot, so it can be
run as 3 ISA/3 PCI or 4 ISA/2 PCI. The board will accept up to 1 MB of
(asynchronous, I believe, and write-back) L2 cache. The board has six
slots for SIMMs, and can use 1/2/4/8/16/32 MB modules. The FDD
controller supports up to a 2.88 MB floppy, and the IDE controller is
the CMD 640 (so use the latest CMD drivers (included) to see all the
second-channel drives). The board also has 2 16550 serial ports and 1
EPP parallel port, and provides support for all the green functions.
The board does not have on-board SCSI, but does have support for the NCR
53c810-based GA-410 NCR 810 PCI SCSI card in the BIOS. The board claims
compatibility with all major operating systems, and I've had a report of
success with it and Warp and Linux.

Setup tips: Use the CMD 640 driver instead of IBM1S506.ADD to
ensure that all IDE drives can been seen and
accessed properly.
Potential problems: None reported so far.

Gigabyte GA-586IP This is Gigabyte's 90/100 MHz
motherboard, and takes one P54CT
running at 60/90 or 66/100 MHz. It has four PCI slots, all of which
allow busmastering, and four ISA slots. It takes either 256 or 512 kB
of L2 cache and up to 768 (!) MB of SIMMs in six 72-pin slots (if you
can get 128 MB strips). The board has the Award flash BIOS (4.50g).
The board has been reported to work reliably under OS/2 (2.1 and Warp),
as well as NT, etc.

Setup Tips: If you are using the Adaptec 2940 SCSI controller
with this board, it is necessary to add the
switches '/A:0 /I' to the basedev line of the
driver. Failure to do so resulted in a consistent
TRAP 3 for one netter. In addition, it is necessary
to set the Int A jumper on the board itself _and_
set the BIOS. Warp full-pack requires the switch
/PCIHW to be added to the device driver line in
config.sys.

Potential problems: In a test in C't they had problems with the ATI
Xpression and the 586IP. Apparently, Gigabyte has
modified PCI slots 0 and 1 (in an unspecified
manner) so that the machine locks up when an ATI
card is installed in one of these slots.
Installing the Xpression in slot 2 is one
workaround.

Intel B486ED This is Intel's 486 PCI board, and
can be outfitted with all the
different Intel CPUs, from the 33 MHz 486 SX to the 100 MHz DX4 (each
CPU gives the boards their own model number; for instance, the 486-DX2
processor-equipped board would be the B486ED8D266). The board may be
upgraded to the Intel P24T processor. It will accommodate up to 256 kB
of cache (128 kB is standard), and 64 MB of 72-pin SIMMs (with or
without parity). The board has IDE and serial/parallel on-board. These
boards use the Saturn II (rev. 4) chipset (I think), and have all the
energy-saving features.

Setup tips: None so far.
Potential problems: None reported so far.

Intel Advanced There are four boards in this
series from Intel, all of which
employ the Triton chipset and AMI BIOS, and will accommodate up to 128 MB
of conventional or EDO RAM. All boards will support the 75 or 90 MHz
chips; some will support the 100 and 120 MHz chips as well. All boards
have on-board busmastering PCI IDE via the Triton chipset, as well as
integrated serial/parallel/game ports. All boards also have 3 dedicated
ISA, 2 dedicated PCI, and 1 shared ISA/PCI slot, so that they can be run
in either 4 ISA/2 PCI or 3 ISA/3 PCI configurations. Specific boards
offer the following features:

Advanced/MN: This board features on-board video (S3/Trio32, up to 2 MB
of DRAM). The board will take up to 256 kB of asynchronous cache. The
low-profile version of this board (MN/LPX) will take up to 512 kB of
cache (still asynchronous). I've had no reports of success with this
board and OS/2.

Setup Tips: None so far.
Potential Problems: None so far.

Advanced/ZP: This board will accommodate up to 256 kB of cache
(asynchronous). It does not have on-board video. I've seen one report
on the net which indicates that, although this board works well with
DOS/Win (what doesn't?), video cards seem to not work well at all with
this board and OS/2. One other report says that a board sounding very
much like this one and using a Diamond Stealth 64 DRAM video card has
worked very well (when used with a 100 MHz Pentium, which this board is
theoretically not supposed to support).

Setup Tips: None so far.
Potential Problems: There may be (as yet unspecified) problems with
this board and certain video cards in OS/2.

Advanced/ZE: This board is very similar to the ZP above, except that it
has one extra dedicated ISA slot and one extra dedicated PCI slot. I
have had no reports of success with this board and OS/2.

Setup Tips: None so far.
Potential Problems: None so far.

Advanced/EV: This board will take all the Pentium processors (up to 120
MHz), and up to 512 kB of synchronous cache. It has built-in audio
(SoundBlaster 16 chipset) and, in the UK and Europe, the S3 Trio64 video
chip. It has four dedicated ISA slots, three dedicated PCI slots, and
one shared ISA/PCI slot. I've had one report of success with this board
and Warp.

Setup Tips: None so far.
Potential Problems: None so far.

Intel AltServer This board is intended for use
primarily in server applications.
It supports either one or two 75 or 90 MHz Pentia, and uses the Neptune
chipset and AMI flash BIOS. Up to 256 kB of (asynchronous) cache may be
used in combination with up to 256 MB of RAM in eight 72-pin strips (EDO
RAM is not supported). The board has an on-board Cirrus Logic 5430
video controller (with 512 kB DRAM expandable to 1 MB) and an Adaptec
AIC7870 fast/wide SCSI controller. There are two dedicated PCI slots,
five dddicated EISA slots, and one shared EISA/PCI slot, and all slots
allow busmastering adapters. When a second (expensive!) 90 MHz CPU is
added to the board with an SMP operating system installed, system
performance is claimed to only increase by 30% according to Intel.

Setup Tips: None so far.
Potential Problems: None so far.

Intel Premiere** The 60/66 (Premiere) and 75/90/100
MHz (Premiere II) Pentium boards
have passed OS/2 certification (with the 66 and 90 MHz CPUs in place,
respectively). These boards are, therefore, highly recommended. The
60 MHz board has been reported to be problematic with SCSI, though
Adaptec now seems to think their AHA-2940 SCSI controller should work
with it, as do BusLogic and QLogic (I have had reports of success with
the QLogic controller but not with the Adaptec or BusLogic). The 60/66
MHz board uses the Mercury chipset, while the 90/100 MHz board uses the
Neptune chipset. All the boards come with 256 kB of L2 cache, can
accommodate up to 128 MB (4x32 MB 72-pin SIMMs) of RAM, and use the AMI
flash BIOS (currently at revision 1.00.12.AX1 for the 90 MHz boards --
other boards have different BIOSes specific to them). All boards have
the NCR SCSI BIOS built in, and have IDE (for ISA and PCI) and I/O ports
on the board. The 60/66 MHz board uses the CMD 10640 chipset for IDE
while the 90/100 MHz board uses the PC Tech RZ1000. The boards can be
run with 4 ISA/3 PCI or 5 ISA/2 PCI slots.

Setup tips: Get (at least) rev. 1.00.10.AX1 of the AMI flash
BIOS (for the 90 MHz boards) to fix problems with
Guaranteed Acess Timing (GAT) and BackMaster 1.1.
If you have an ATI card and an intermal modem, make
sure you turn off intelligent remapping of the COM
ports to avoid conflicts with the ATI card and
COM4. If you are using an NCR SCSI controller,
you'll have to set IRQ9 to 'used by ISA card'
during the install to get the install to work
(under 2.1/2.11). If you're running a SCSI drive
as your boot drive, turn off the drive C: timeout
for a faster boot.

WARNING *** WARNING *** WARNING *** WARNING *** WARNING *** WARNING
The following setup information is provided with ABSOLUTELY NO WARRANTY
of ANY KIND. In no manner whatsoever shall I (Patrick Duffy, the author
of the PCI motherboards list) be held responsible for damage of any sort
caused by application of said information.

That said, here's how to make your 90 MHz Premiere II board run at 100 MHz:

Move the "reserved" jumper on the board to pins 1 & 2 (from pins 2 & 3,
I think). This causes the Pentium to run at 100 MHz instead of 90.
WARNING *** WARNING *** WARNING *** WARNING *** WARNING *** WARNING

Potential problems: See the PCI chipset list and the PCI IDE list for
descriptions of the Neptune chipset and IDE drive
problem, The SMC chip used to control the serial
ports should have the letters "GT" after it for
trouble-free communications.

DEALS OF THE WEEK: Intel P90 CPU: Computime $599 (619) 268-8856
Pixel $618 (408) 929-7218
Computer Integration Tech $639
(800) 730-2983
Roland Baker $645
ba...@ocf.berkeley.edu

Intel Premiere II P90 Motherboards:
Roland Baker $335
ba...@ocf.berkeley.edu
Spire Tech $407 sp...@teleport.com
Tony $360 (818) 281-8628

J-Bond PCI400C-A This board will use up to a DX2-66
CPU, and has a spot for a Pentium
Overdrive chip (whatever those really are). It has rev. 2 of the Saturn
Chipset (the latest boards do), and the Phoenix BIOS rev. 1.03 (the
board reported has a BIOS dated April 2, 1994). The board has a
built-in NCR SCSI controller (presumably based on the 53c810, though
this was not indicated). It boots OS/2 successufully, though there
seems to be long delays before bootup when a Quantum LPS540S hard drive
is used (but not a Seagate). (This seems to be a problem with that
particular Quantum drive.) The board takes up to 4 72-pin 36-bit SIMMs,
installed in identical pairs. The board will take either three PCI and
four ISA or two PCI and five ISA cards (one slot is shared). The board
has no on-board I/O at all (except for the SCSI controller). Warp beta
II has apparently installed fairly painlessly on this board, though
there are random lockups which seem to be due to a design shortcoming of
the motherboard. I wouldn't recommend this board.

Setup tips: Disable the external cache for reliable SCSI
operation.
Potential problems: See the PCI chipset list for a description of the
problems with the Saturn I (rev. 2) chipset.

J-Bond PCI500C-A This board is, from a report I've
received, no better than the 486
board above. Apparently the only way to make _this_ board stable (the
test configuration had a 66 MHz Pentium, 512 kB cache, Phoenix BIOS
1.03, shadow RAM enabled, Mercury chipset, and the NCR 53c810 controller
(built in) attached to a Quantum SCSI HD) is to disable the on-chip 16
kB cache. This, of course, makes the Pentium _really_ slow. New boards
may have this problem fixed (whatever's causing it), but I have no way
of knowing. If you're still curious, the board has 4 PCI and 4 ISA
slots, and can accommodate up to four 72-pin SIMMs.

Setup tips: Disable the CPU cache for reliable operation.
Potential problems: Motherboard problems could be caused by:
- Early BIOS (Feb. 02/94, rev. 1.03)
- Early Mercury chipset (not sure)
- Bad CPU or poor cooling

Lexar LXM-510 This motherboard will take all the
Intel 486 CPUs and has a spot for a
Pentium Overdrive socket (though to use the 3.3 V processors the "Model
99 Regulator" must be purchased). Early revisions of the board had a
separate connector for 3.3 V power to the PCI sockets; this has been
replaced with a separate voltage regulator in later versions. It will
take up to 128 MB of RAM in 8 30-pin SIMM sockets, and up to 512 kB of
cache. It uses either the Award or AMI flash BIOSes, and has 2 VL, 2
ISA, and 2 PCI slots. It uses the IMS (Integrated Micro Solutions)
chipset. The board also has all the standard on-board super I/O and a
mouse port. It has been tested and found to be compatible with DOS/Win
and Win/NT (no mention of OS/2). I have had a report of success with
this board and DOS/Win and Linux.

Setup tips: The manual, while apparently nicely printed, is
poorly written. This may cause problems when
configuring the many jumpers on the board.
Potential problems: None so far.

Microgram ??? This motherboard will run at 25 or
33 MHz, has 5 ISA and 3 PCI slots,
and a Phoenix BIOS. The board will also accommodate up to 128 MB of
RAM, in 4 72-pin SIMM slots. Everything up to a DX2-66 can be put in,
as can a P24T into the available ZIF socket. The new DX4s will not work
because the board does not support 3.3V. (Though I suppose you could
get a 5->3.3V adapter if you _really_ wanted to.) This board apparently
runs OS/2 reasonably well, though there seem to be random lockups at
times. There were no SCSI devices on the board, and apparently with an
old WD90C11 video card things were quite erratic. I don't know what
chipset this board uses, but from the range of processors I'm guessing
Saturn, so make sure it's rev. 4 if you want SCSI (the board tested had
rev. 1 of the Saturn chipset).

Setup tips: None so far.
Potential problems: See the PCI chipset list for a description of the
problems with the Saturn I (rev. 2) chipset.

Micronics M4Pi This is Micronics' 486 PCI
motherboard. It will take
everything up to a DX4 (with a ZIF socket for a Pentium overdrive), and
supports 3.3V. It is (feature-wise) identical to the M5Pi board
below, except that it uses the Intel 82420 PCIset PCI chipset (the
Saturn chipset), and has an extra dedicated ISA slot.

Setup tips: None so far.
Potential problems: None reported.

Micronics M5Pi This particular board takes
either a 60 or 66 MHz Pentium chip,
and has a ZIF socket for future upgrades. Like many of the other
Pentium motherboards here, this has 256/512 kB of (write-back) cache,
and will accommodate up to 128 MB (4x32 MB SIMMs) of RAM. The PCI
chipset used is the Intel 82430 PCIset (don't know it's common name,
though I suspect that this is the Mercury chipset), and it uses a
Phoenix Flash-BIOS. Like the Intel boards, this board can be run with 4
ISA and 3 PCI slots or 5 ISA and 2 PCI slots.

Setup tips: None so far.
Potential problems: None reported.

Micronics M54pi* This is the 90 MHz Micronics
motherboard for OS/2. Presumably
it has the Neptune chipset. The board reported has Phoenix BIOS V4.04-N
08. With this BIOS, the board is reported to work very well with the
Adaptec 2940, though earlier versions of the BIOS had problems with
disk-intensive programs crashing. Make sure, therefore, that you get at
least this revision of the Phoenix flash BIOS when buying your
motherboard, or at least that you can upgrade to it. I have now had
several reports of success with this board and various flavours of OS/2
(all using the 2940 controller).

Setup tips: Make sure you have the latest revision of the BIOS
for your best chance at trouble-free operation.
Potential problems: None, with the latest BIOS.

SOYO 80486 As its name might suggest, this is
a 486 board which accommodates all
the different 80486 chips and the overdrive series as well. It has all
the green features, though the BIOS type is unspecified. The board has
four ISA (two of which are VL) and four PCI slots (all of which allow
busmastering). The board will take up to 128 MB of RAM, and will
accommodate cache RAM in increments of 32 kB. It is reported to work
well with Warp.

Setup Tips: None so far.
Potential Problems: None so far.

SuperMicro P55* This motherboard is based around
the Intel Triton chipset, and as
such will support the 75, 90, 100, and 120 MHz CPUs. Apparently when
(if) the 133, 150, and 180 MHz P5s are released it will support these as
well. The board itself comes in two types: CWA, which has an
asynchronous cache, and CWS, which has a pipelined burst synchronous
cache. Either board can have 256 or 512 kB of cache. The board has 3
ISA slots, 3 PCI slots, and one shared ISA/PCI slot, all of which allow
busmastering. The board has 4 SIMM sockets, for up to 128 MB of RAM, in
either 60/70 ns fast page-mode or EDO. The board has EIDE support
through mode 4, but that's about it for I/O, as it has no built-in
serial/parallel/game or floppy. In return for this the board is
relatively inexpensive, costing less than $US 1000 with the 100 MHz
pentium. I've had two reports of success with this board and OS/2 Warp
(full-pack, CWS configuration, EDO RAM, 100 MHz), as well as DOS/Win,
NT, and SCO Unix (75 MHz), though apparently with SCO Unix the SCSI
controller fails after a warm boot and must be reset.

Setup Tips: None so far.
Potential Problems: None so far.

TMC PCI48PG4 This is a combination VL/ISA/PCI
board which uese the Opti chipset.
It uses the Opti PCI IDE controller (82C621), and the SMC chip for
serial/parallel/floppy I/O. It takes all the different 486 processors
and the P24D overdrive chip. The board has 2 PCI slots, one shared
PCI/ISA, 2 ISA, and 2 VL slots. The board will accommodate up to 128 MB
RAM (in four slots), though apparently if the first two slots are filled
the second two must be filled as well or the board will not work. The
board takes up to 256 kB cache and uses the Award or AMI BIOSes, though
it is reported to work only with the AMI (WinBIOS) BIOS. Apparently
also the 7/25/94 WinBIOS would not allow the computer to boot with a
Stealth 64 VRAM or a Stealth 64 Video VRAM installed. Also, the board
ran very slowly with all four SIMM sockets filled and this BIOS.
Downgrading to the 12/15/93 version of the BIOS fixed these problems,
but limited control over the peripheral I/O (there were few options
present in the BIOS setup to allow it).

Setup tips: Until AMI comes out with a later version (than
7/25/94) of the WinBIOS, use 12/15/93 for best
results.
Potential problems: None, with the 12/15/93 BIOS.

TMC PCI54IT This board is based around the
Intel Triton chipset and the Award
BIOS (revision 4.50GP, 95/04/20). It is plug and play 1.0a-compliant.
It will take at least up to a 90 MHz Pentium and has four 72-pin SIMM
slots. Support for IDE (type unspecified), floppy, and
serial/parallel/game I/O is built in to the motherboard. The board is
reported to work well with Warp.

Setup Tips: None so far.
Potential Problems: None so far.

TMC PCI54PV3 This motherboard uses up to a 90 MHz
Pentium chip and has the Opti "Viper"
chipset in combination with the Award BIOS. It has 3 PCI slots, 4 ISA
slots, and all the usual serial/parallel/game ports. The board will take
up to four 72-pin SIMMs. Support for IDE and floppy drives is built in,
though what type of IDE is unspecified. The board is reported to work
well, though somewhat slower than the PCI54IT above, with Warp.

Setup Tips: None so far.
Potential Problems: None so far.

TMC PCI58PL This board will take either a 60 or
66 MHz Pentium, and up to 128 MB of
(parity only) RAM. The memory configuration is unique, though, in that
the board has four sockets for 30-pin SIMMs and three for 72-pin SIMMs.
The 30-pin sockets (all four) comprise bank 0, while the 72-pin sockets
are banks 0, 1, and 2, so that if you use 30-pin SIMMs you'll only have
two sockets left over for 72-pin SIMMs. The board will take 1Mx9, 4Mx9,
16Mx9, 256kx36 (total 1 MB), 512kx36, 1Mx36, 2Mx36, 4Mx36, 8Mx36, and
16Mx36 SIMMs for a maximum of 192 MB. It will take up to 512 kB of
cache, and comes with 256 kB. The board has 2 ISA, 1 VL, and 3 PCI
slots, all dedicated, and one shared PCI/VL slot. All PCI slots allow
busmastering, and one of the VL slots does. It has no built-in I/O
support for floppy drives or serial/parallel ports. The board uses the
OPTI 82C822, 82C571, and 82C572 chipset and the Award (4.50G) BIOS. The
board is reported to work well with OS/2 Warp and DOS/Win.

Setup Tips: None so far.
Potential Problems: The version of the Award BIOS which comes with this
board has a bug. Apparently pressing [F1] for help
when in the PCI configuration screen does not
produce any help.

UMC UMC88 This board takes a 486 CPU (type
unspecified, but presumably all of
them), and has all the latest green features. The board is reported to
work well with OS/2 (version unspecified), with the single exception
that it will not allow OS/2 to boot from a floppy drive for one user.
Another person reports that OS/2 will boot from floppy, but that the
board is unstable, and that OS/2 experiences random lockups. Replacing
the two 8 MB SIMMs with one 16 MB SIMM fixed the problem, which suggests
that the reported problems could be due to a bad SIMM. Replacing the
SCSI disk also helped (two bad hardware components at one go?)
Apparently too the board/video card combination does not seem to work
well with FeelX, causing random video corruption when FeelX is installed
(the card is a Cirrus Logic 543X).

Setup Tips: None so far.
Potential Problems: OS/2 might not boot from the floppy drive with this
board.

There's what I know. Please E-Mail suggestions/corrections and I'll
post again.
--
Patrick Duffy, du...@theory.chem.ubc.ca

"Never fry bacon when you're naked."

tray...@pcix.com

unread,
Jul 25, 1995, 3:00:00 AM7/25/95
to
In <3uvcsb$1...@nntp.ucs.ubc.ca>, du...@theory.chem.ubc.ca (Pat Duffy) writes:
> Document: PCI Motherboards for OS/2 list
> Maintainer: Patrick Duffy, du...@theory.chem.ubc.ca
>Last Revision Date: July 23, 1995
> Archived at: ftp.netcom.com, in directory /pub/ab/abe/
> Web pages: http://warp.eecs.berkeley.edu/os2/workbench/work.htm

[stuff deleted]

> Intel Premiere** The 60/66 (Premiere) and 75/90/100
> MHz (Premiere II) Pentium boards
>have passed OS/2 certification (with the 66 and 90 MHz CPUs in place,
>respectively). These boards are, therefore, highly recommended. The
>60 MHz board has been reported to be problematic with SCSI, though

[more deleted]

>on the board. The 60/66 MHz board uses the CMD 10640 chipset for IDE

Not quite true. I have a 60/66 MHz Premiere board with the PC Tech chip.
I looked after the Powerquest dmatest program reported errors. Also,
according to IBM on compuserve the problem is with just the PC Tech chip.

>while the 90/100 MHz board uses the PC Tech RZ1000. The boards can be
>run with 4 ISA/3 PCI or 5 ISA/2 PCI slots.

Terry Raymond

Pat Duffy

unread,
Jul 27, 1995, 3:00:00 AM7/27/95
to
In article <3v8a4g$5...@tuegate.tue.nl>,
Peter Kruijt <pet...@wfw.wtb.tue.nl> wrote:

>Also the 90 MHz Intel boards have this DMA bug.

Here's what causes it, straight from IBM:

2) PC-Tech RZ-1000 [confirmed/software fix implemented in Warp fixpack 5]

Produces accelerated IRQ. IBM1S506.ADD actually tolerates this.
However, when the driver reads 1x7 during interrupt processing to
clear the interrupt, the chipset puts the status in its read-ahead
FIFO resulting in data corruption. Intel used this chip on their
motherboards which they sold to Gateway, AT&T, IBM and others.

This feature is not mentioned in its chipset errata.

Peter Kruijt

unread,
Jul 27, 1995, 3:00:00 AM7/27/95
to
In article <26KJPh$0...@zoe.pcix.com> tray...@pcix.com (Terry Raymond) writes:
>
>[stuff deleted]

>
>> Intel Premiere** The 60/66 (Premiere) and 75/90/100
>> MHz (Premiere II) Pentium boards
>>have passed OS/2 certification (with the 66 and 90 MHz CPUs in place,
>>respectively). These boards are, therefore, highly recommended. The
>>60 MHz board has been reported to be problematic with SCSI, though
>
>[more deleted]

>
>Not quite true. I have a 60/66 MHz Premiere board with the PC Tech chip.
>I looked after the Powerquest dmatest program reported errors.
>
[more deleted]
Also the 90 MHz Intel boards have this DMA bug. Highscreen/Vobis (relatively
large retailer in Germany/The Netherlands) sells their `home brand' Pentiums
with (presumably, I'm not entirely sure) the Premiere II motherboard. Setting
DMA prefetch buffers to disabled in the CMOS setup resolves this problem,
(as suggested by PowerQuest), although it is not a real solution (replacement
is the REAL solution)

just my 0.02/0.02=256 cents worth,
--
Peter Kruijt (pet...@wfw.wtb.tue.nl) Eindhoven University of Technology
Department of Mechanical Engineering PO Box 513
5600 MB Eindhoven

Pat Duffy

unread,
Jul 30, 1995, 3:00:00 AM7/30/95
to
Document: PCI Motherboards for OS/2 list
Maintainer: Patrick Duffy, du...@theory.chem.ubc.ca
Last Revision Date: July 26, 1995

Archived at: ftp.netcom.com, in directory /pub/ab/abe/
Web pages: http://warp.eecs.berkeley.edu/os2/workbench/work.htm

This is the PCI motherboards for OS/2 list. For detailed descriptions

Useful information:

Intel Premiere** The 60/66 (Premiere) and 75/90/100


MHz (Premiere II) Pentium boards
have passed OS/2 certification (with the 66 and 90 MHz CPUs in place,
respectively). These boards are, therefore, highly recommended. The
60 MHz board has been reported to be problematic with SCSI, though

Adaptec now seems to think their AHA-2940 SCSI controller should work
with it, as do BusLogic and QLogic (I have had reports of success with
the QLogic controller but not with the Adaptec or BusLogic). The 60/66
MHz board uses the Mercury chipset, while the 90/100 MHz board uses the
Neptune chipset. All the boards come with 256 kB of L2 cache, can
accommodate up to 128 MB (4x32 MB 72-pin SIMMs) of RAM, and use the AMI

flash BIOS (currently at revision 1.00.13.AX1 for the 90 MHz boards --


other boards have different BIOSes specific to them). All boards have
the NCR SCSI BIOS built in, and have IDE (for ISA and PCI) and I/O ports

on the board. All boards use the PC Tech RZ1000 for IDE support. The


boards can be run with 4 ISA/3 PCI or 5 ISA/2 PCI slots.

Setup tips: Get (at least) rev. 1.00.10.AX1 of the AMI flash

"Old people don't need companionship; they need to be isolated and studied to
see if they contain any nutrients that we might need." -- Homer J. Simpson

tray...@pcix.com

unread,
Jul 30, 1995, 3:00:00 AM7/30/95
to
In <3v8vn4$m...@nntp.ucs.ubc.ca>, du...@theory.chem.ubc.ca (Pat Duffy) writes:
>In article <3v8a4g$5...@tuegate.tue.nl>,
>Peter Kruijt <pet...@wfw.wtb.tue.nl> wrote:
:
::Also the 90 MHz Intel boards have this DMA bug.
:
:Here's what causes it, straight from IBM:

:
:2) PC-Tech RZ-1000 [confirmed/software fix implemented in Warp fixpack 5]

Do you mean that systems with fixpack 5 installed should pass the
dmatest? If so, then there is more too it because I have fixpack 5 installed
and my system failed the test.

:
: Produces accelerated IRQ. IBM1S506.ADD actually tolerates this.


: However, when the driver reads 1x7 during interrupt processing to
: clear the interrupt, the chipset puts the status in its read-ahead
: FIFO resulting in data corruption. Intel used this chip on their
: motherboards which they sold to Gateway, AT&T, IBM and others.
:
: This feature is not mentioned in its chipset errata.

Terry Raymond

Pat Duffy

unread,
Jul 30, 1995, 3:00:00 AM7/30/95
to
In article <26NLyj$0...@zoe.pcix.com>, <tray...@pcix.com> wrote:

>:2) PC-Tech RZ-1000 [confirmed/software fix implemented in Warp fixpack 5]
>
>Do you mean that systems with fixpack 5 installed should pass the
>dmatest? If so, then there is more too it because I have fixpack 5 installed
>and my system failed the test.

Hmmmm... It _should_ pass the test with fixpack 5 installed. What version of
the Neptune chipset do you have?

0 new messages