Bug in Via Nehemiah C3 stepping 4 CPU?

[Jay]

Hello folks,

I was wondering if anyone has seen this problem, and whether there's a
work around (microcode update?):

We have two [seemingly] identical embedded processor boards that use VIA's
1.0GHz Nehemiah C3 CPU. One has no issues running win2k, the other will
not run the 16 bit application "wowexec.exe" - the error "NTVDM has
generated errors and will be closed by windows" occurs.

Upon closer examination, the failing board has a stepping 4 processor, and
the working board has a stepping 8 processor

Running Microsoft's debugger reveals the following:

Symbol search path is: SRV*C:\symbols*http://msdl.microsoft.com/download/symbols;C:\symbols

Microsoft (R) Windows Debugger Version 6.3.0017.0
Copyright (c) Microsoft Corporation. All rights reserved.

*** wait with pending attach
Symbol search path is: SRV*C:\symbols*http://msdl.microsoft.com/download/symbols;C:\symbols
Executable search path is:
ModLoad: 0f000000 0f0a1000 C:\WINNT\system32\ntvdm.exe
ModLoad: 77f80000 77ffd000 C:\WINNT\system32\ntdll.dll
ModLoad: 7c570000 7c628000 C:\WINNT\system32\KERNEL32.dll
ModLoad: 7c2d0000 7c332000 C:\WINNT\system32\ADVAPI32.dll
ModLoad: 77d30000 77da1000 C:\WINNT\system32\RPCRT4.DLL
ModLoad: 77f40000 77f7e000 C:\WINNT\system32\GDI32.dll
ModLoad: 77e10000 77e75000 C:\WINNT\system32\USER32.DLL
ModLoad: 732e0000 73305000 C:\WINNT\system32\shim.dll
ModLoad: 23000000 23056000 C:\WINNT\AppPatch\AcLayers.DLL
ModLoad: 69b00000 69b07000 C:\WINNT\system32\NTVDMD.DLL
ModLoad: 0ffc0000 10000000 C:\WINNT\system32\WOW32.dll
ModLoad: 782f0000 78535000 C:\WINNT\system32\SHELL32.dll
ModLoad: 70a70000 70ad5000 C:\WINNT\system32\SHLWAPI.dll
ModLoad: 78000000 78045000 C:\WINNT\system32\msvcrt.dll
ModLoad: 71710000 71794000 C:\WINNT\system32\COMCTL32.dll
ModLoad: 76b30000 76b6e000 C:\WINNT\system32\comdlg32.dll
ModLoad: 77820000 77827000 C:\WINNT\system32\VERSION.dll
ModLoad: 759b0000 759b6000 C:\WINNT\system32\LZ32.DLL
(378.374): Access violation - code c0000005 (!!! second chance !!!)
eax=00000000 ebx=0f095bc4 ecx=00004212 edx=00000000 esi=00020000 edi=fffdffff
eip=0f03e255 esp=0105fc40 ebp=0105ff40 iopl=0 nv up ei pl zr na po nc
cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000 efl=00000246
ntvdm!FastLeavePm+0x35:
0f03e255 cf iretd

Has anybody else had problems with the stepping 4 C3 processors?

Thanks for any info!

Best Regards,

Jay Walling
Diagnostic Engineer
(e-mail address removed)
Comark Corporation
93 West St.
Medfield, MA 02052
http://www.comarkcorp.com

 

[The little lost angel]

We have two [seemingly] identical embedded processor boards that use VIA's
1.0GHz Nehemiah C3 CPU. One has no issues running win2k, the other will
not run the 16 bit application "wowexec.exe" - the error "NTVDM has
generated errors and will be closed by windows" occurs.

I'm not sure about these things, but have you tried running some
memory tests or swapping their RAM modules? I've seldom see actual
processor failures, but memory quite often.

--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me
Standard HTML, SHTML, MySQL + PHP or ASP, Javascript.
If you really want, FrontPage & DreamWeaver too.
But keep in mind you pay extra bandwidth for their bloated code

 

[Robert Redelmeier]

We have two [seemingly] identical embedded processor boards that use VIA's
1.0GHz Nehemiah C3 CPU. One has no issues running win2k, the other will
not run the 16 bit application "wowexec.exe" - the error "NTVDM has
generated errors and will be closed by windows" occurs.

I'm not sure about these things, but have you tried running
some memory tests or swapping their RAM modules? I've seldom
see actual processor failures, but memory quite often.

Always a good idea, but a repeatable error usually
is software not hardware. Only if the software exec
history is identical is bad memory cell likely.

Ms-Win2k is "complicated" enough that even running
`copy CMD.EXE c.EXE` or launching & killing an app
will change memory usage (page allocation).

-- Robert

 

[Jay]

We have two [seemingly] identical embedded processor boards that use VIA's
1.0GHz Nehemiah C3 CPU. One has no issues running win2k, the other will
not run the 16 bit application "wowexec.exe" - the error "NTVDM has
generated errors and will be closed by windows" occurs.

I'm not sure about these things, but have you tried running
some memory tests or swapping their RAM modules? I've seldom
see actual processor failures, but memory quite often.

Thanks, but in this situation, we've done all these thing already.

Always a good idea, but a repeatable error usually
is software not hardware. Only if the software exec
history is identical is bad memory cell likely.

It's not. We're now up to 7 boards - four have stepping 8 and all work,
three of them are stepping 4 and all fail. Pretty conclusive, I should
think. I've examined the rest of the PCI devices using a PCI utility to
check the device IDs and revision level of the chips, and all 7 boards are
identical in that respect. Revision level of the raw circuit board is also
identical. I've even gone as far as examining all the SMD components
(resistor values, etc.) and there is no difference there, either. BIOS
code and BIOS setups - identical. There is no other differences that I can
discern.

The only conclusion I can thing of is that the BIOS needs updating; maybe
for a microcode update, but according to the manufacturer, it's the latest
BIOS. It would be nice to be able to swap a stepping 8 CPU into a board
with a stepping 4 CPU, but, unfortunately, these are BGA devices and are
soldered onto the board. Does anyone know if VIA publishes errata for
their CPU chips like Intel does? If they do, does it require an NDA? I was
unable to find any on their site.

--
Jay Walling
Diagnostic Engineer
(e-mail address removed)
Comark Corporation
93 West St.
Medfield, MA 02052
http://www.comarkcorp.com

 

[Robert Redelmeier]

It's not. We're now up to 7 boards - four have stepping 8 and
all work, three of them are stepping 4 and all fail. Pretty
conclusive, I should think. I've examined the rest of the

Yep, that sounds very conclusive of some systemic problem.

PCI devices using a PCI utility to check the device IDs and
revision level of the chips, and all 7 boards are identical

Check the configuration space to make sure the same values
got written to the Northbrige (&southbridge, if any). `lspci`
under Linux IIRC.

The only conclusion I can thing of is that the BIOS needs
updating; maybe for a microcode update, but according to the

Maybe, if the C3 uses loadable microcode. Intel definitely
does after their Pentium FDIV fiasco. I'm not sure about
AMD or the others.

soldered onto the board. Does anyone know if VIA publishes
errata for their CPU chips like Intel does? If they do, does
it require an NDA? I was unable to find any on their site.

Sorry, I don't know but would be interested in the answer.

-- Robert

 

[Keith R. Williams]

Maybe, if the C3 uses loadable microcode. Intel definitely
does after their Pentium FDIV fiasco. I'm not sure about
AMD or the others.

Apparently AMD does.

http://www.realworldtech.com/forums/index.cfm?action=detail&PostNum=
2527&Thread=1&entryID=35446&roomID=11

 

[lesterz]

I 've got the same pb as you
My PC is a laptop with a via C3 nehemiah at 1 GHz. My OS is a win2k SP4
and every ntvdm software crashes
I've perhaps found an explainationto our problem on microsoft site
http://support.microsoft.com/default.aspx?scid=kb;en-us;817901#appliesto
On the CD given with the PC I've got a patch file (ntvdm.exe) but only for
win 2k, win2k SP2, win2k SP3 but not for SP4
Can you help where to find this file!!!

Sorry for my english