See attached.

Pentium 200 MMX
64MB Ram
i430 TX Motherboard
Matrox Mill 2 8mb
Voodoo 2 12mb
Sound Blaster 32

red-ray wrote on 2019-08-22, 11:54:

mwdmeyer wrote:

Voodoo 2 12mb

Thank you, after checking the save file all looks to be as it should be. Given you have both a Millennium II and 3dfx Voodoo2 what does the 3dfx Voodoo2 do? I have just started adding 3dfx support and wondering, does/can it have a display attached?

No it just offers 2d pass through and full screen 3d.

I did manage to get a desktop working with Windows 2000 and a beta driver which was pretty cool, no idea how it worked!

mwdmeyer wrote:

I did manage to get a desktop working with Windows 2000 and a beta driver which was pretty cool, no idea how it worked!

Thank you, was this on the same system? If so and you run SIV on Windows 2000 it should also report such as the Disk SMART + DIMM SPD information. I have just updated the first post to recommend using NT.

I installed the vcredist you provided on my T2130CS (486DX4-75, 24MB RAM, 960K Chips&Technologies VGA) with 95A, and both CPU-Z for Win95 and SIV32L cause the machine to completely lock up. I wonder what it's angry about? DOS mode utilities like SpeedSys work fine. Odd. I will thank you for providing the redistributable though -- I wasn't able to find an appropriate version on my own that worked with 95 A. I needed it for 7-Zip 9.20.

Sedrosken wrote:

both CPU-Z for Win95 and SIV32L cause the machine to completely lock up

Thank you for trying out SIV, apologies for the lockup and I am pleased to hear 7-ZIP now works for you. Did you try SIV on your Pentium Pro NT 4.0 SP6a system and was it OK on that? Did it report the 3Dfx Voodoo Banshee clock speeds?

I would like to try and fix SIV and to do this I need your help in finding out where the lockup is happening. If the MSVC 6 remote debugger could be used this would be easy, but given what you said about the redistributable guess you don't have this.

From a command window please do SIV32L -DBGINI -DBGTIM -DBGPCI > SIV_DBGOUT.log | more or just SIV32L -DBGINI -DBGTIM -DBGPCI and post SIV_DBGOUT.log so I can see what SIV is doing when the lockup happens.

I can test it on my Pentium 133 430VX system, I need to dig it up (I'll see if I can report back on 1st September, I'm not currently at home right now...)

FWIW, SIV32L and SIV32X (either one) crashes on my Win2000 socket AM2+ system when it gets to "UEFI info" if I try to save a report.

Vynix wrote:

I can test it on my Pentium 133 430VX system, I need to dig it up (I'll see if I can report back on 1st September, I'm not currently at home right now...)

Thank you and no real rush.

bakemono wrote:

FWIW, SIV32L and SIV32X (either one) crashes on my Win2000 socket AM2+ system when it gets to "UEFI info" if I try to save a report.

Thank you for checking out SIV and I am rather surprised [uefi-info] is there at all on a Windows 2000 system. What happens if you do Menu->File->Save->Save Selected and deselect [uefi-info] before pressing [ Start ] ?

Please will you also do Menu->File->Save->Save System ROM and post the .ROM file in addition to the .TXT + .DMI files? My guess is that SIV is crashing while scanning the contents of the System ROM.

I tried it on my ComputerLab ST5500 thin client. Everything seems to work, but SIV disagrees with CPU-Z on whether it's a DDR1 or DDR2 system.

I'm pretty sure CPU-Z is correct here.

This is a very compact thin client; I haven't opened it up to verify this, because that requires breaking plastic tabs, but I'm pretty sure the RAM is soldered onto the mainboard (no SODIMMs.) Note that the CPU-Z 1.89.3 x32 beta in the other thread also has trouble reading this system's RAM info but 1.86 has no problem.

Also, the CPU is reported as a "VIA Eden 500MHz" running at 1GHz (5x200). I think this is a problem with the CPUID as it happens in everything, including CPU-Z, Windows's device manager, and Linux's /proc/cpuinfo.

Full report here:

xjas wrote:

CPU is reported as a "VIA Eden 500MHz" running at 1GHz (5x200). I think this is a problem with the CPUID as it happens in everything, including CPU-Z, Windows's device manager, and Linux's /proc/cpuinfo.

1. Thank you for checking out SIV and your excellent post. I have adjusted SIV, there is now a 5.41 Beta-17 and I noted:
2. You are correct in that it should be DDR2 and Beta-17 should get all the memory timings correct. Note CPUZ fails to report tRFC.
3. SIV failed to read the DIMM SPD data and Beta-17 should, does it?
4. If you hover the mouse over the date/time on the 2nd line SIV will report how long it took to read the SPD data which I suspect will be rather faster than CPUZ, is it?
5. Looking @ [Machine] the BIOS reports 1GHz, comparing the SIV CPU benchmarks for my VIA C3-M (Nehemiah, C5P) @ 800 MHz and your system I am pretty sure it's running @ 1GHz.

Please will you try Beta-17 and post both SIV_PHOBOS.txt + SIV_PHOBOS.dmi. The .DMI is a binary dump of the BIOS which contains the data reported by the [Machine] panel and SIV works better in test mode when I have it.

No problem! I tried the new build and it seems to be working fine; it detects DDR2 & reads the SPD correctly. SPD read time was 0.766s (CPU-Z 1.89 hangs there for at least 15s.)

Updated report (including .DMI):

I've done a couple other machines that might be of interest but I'll make a new post for them shortly. BTW big thanks for still supporting these old machines with your program, it's a really useful tool to keep around.

The other two I've done so far are my Shuttle XPC (Geode NX1750+Geforce3) and K6-2/500+Voodoo3+PowerVR, both running Windows 98 (so no detailed machine info, sorry...) Here's what I got:

K6-2+Voodoo3:

Everything seemed to read fine on this one, including identifying & detecting the clocks on my V3 correctly, although it doesn't seem to get much more else from the card.

The PowerVR PCX2 (Matrox M3D) in the system was not detected as a "GPU" but does show up in the PCI device listing. This is a dedicated 3D accelerator that does not have a video output, it communicates with the primary card (V3 in this case) over the PCI bus. Not sure how SIV handles multiple GPUs but just noting this in case it's useful.

System report:

Shuttle XPC (Geode+Geforce 3):

Note: this system is running 98SE with KernelEX installed (the K6-2 doesn't have it.) Not sure that makes any difference. This is also a less "pristine" system meaning I've had it for a while & have swapped around graphics & sound cards/drivers a bit.

Everything seems correct. It identifies the Geode NX1750 CPU as a "Mobile Athlon Thoroughbred", which is more-or-less the same thing... I'm not sure it's even possible to differentiate this model of Geode. The BIOS POST calls it an "Athlon 4 XP-M."

It doesn't identify the GPU on my GeForce 3 Ti200, although it does report it as an ASUS V8200T2, which is correct. It can't seem to read much info from the card.

When saving a report I get a crash at Section 103 [menus]. This happens every time, including on clean boot.

System report:

Let me know if you need any further info from these.

I also have another thin client running a 533MHz VIA C3 on a variant of the PLE133 chipset. If it helps I can try a run on that system too.

xjas wrote:

BTW big thanks for still supporting these old machines.

Thank you for checking things out and all now looks to be OK. The SIV start-up time is rather longer than usual which I think is down to the speed of the D: flash drive.

BTW the system seems to have a SO-DIMM with the AD2SHH1GS1WB-6EGE part number. When the memory is soldered onto the board I have never seen any SPD data. It does not report the manufacturer.

I suspect if you start CPUZ when SIV is active then SIV may whinge about this and if you watch Menu->Help->Lock Handle you will notice CPUZ holds the Global\Access_SMBUS.HTP.Method for an excessive time (> 1 second). Given SIV can read the SPD data in 0.766 seconds then I can't see why CPUZ can't.

I will look forward to seeing how SIV does on your other systems and hope there are no lockup issues.

red-ray wrote:

Thank you for checking things out and all now looks to be OK. The SIV start-up time is rather longer than usual which I think is down to the speed of the D: flash drive.

^^ yeah, drive D is a really slow USB stick. C: is only a 1GB IDE flash & is mostly taken up by Win2K so I make do.

xjas wrote:

Shuttle XPC (Geode+Geforce 3): […]
Show full quote

Shuttle XPC (Geode+Geforce 3):

1. Note: this system is running 98SE with KernelEX installed (the K6-2 doesn't have it.) Not sure that makes any difference. This is also a less "pristine" system meaning I've had it for a while & have swapped around graphics & sound cards/drivers a bit.
2. Everything seems correct. It identifies the Geode NX1750 CPU as a "Mobile Athlon Thoroughbred", which is more-or-less the same thing... I'm not sure it's even possible to differentiate this model of Geode. The BIOS POST calls it an "Athlon 4 XP-M."
3. It doesn't identify the GPU on my GeForce 3 Ti200, although it does report it as an ASUS V8200T2, which is correct. It can't seem to read much info from the card.
4. When saving a report I get a crash at Section 103 [menus]. This happens every time, including on clean boot.

1. Thank you for the report
2. OK
3. Looking @ Menu->Hardware->CPUID->CPU-0 I expect this is as good as is possible, does CPUZ report the same?
4. For NVidia GPUs SIV uses NVAPI and your GPU is too old for NVAPI to support it. If it was an NT system SIV should report the GPU chip.
5. I suspect this is down to a bug in KernelEX as I have never seen this before. Do Menu->File->Save->Save Selected and deselect [menus] before pressing [Start].
6. If you uninstall KernelEX do Menu->Help->View Menus and/or saving [menus] work OK?
7. Which version of KernelEX do you have and where can I get it from so I can see what happens on one of my systems and try to adjust SIV.

xjas wrote:

Everything seemed to read fine on this one, including identifying & detecting the clocks on my V3 correctly, although it doesn't seem to get much more else from the card.

The PowerVR PCX2 (Matrox M3D) in the system was not detected as a "GPU" but does show up in the PCI device listing. This is a dedicated 3D accelerator that does not have a video output, it communicates with the primary card (V3 in this case) over the PCI bus. Not sure how SIV handles multiple GPUs but just noting this in case it's useful.

Thank you for the report and this is the first time I have see SIV report Voodoo 3 clocks 😎 AFAIK there is nothing else that could be reported.

The NEC PowerVR PCX2 3D Accelerator is an Other Multimedia so is not really a GPU unlike the 3Dfx Voodoo3 which is VGA Controller.

On NT systems SIV does support multiple GPUs, 8 by default and -GPUS=<N> can be used to support more. Multiple should work on 98 and I got as below when I tried it, notice how SIV does rather better on 2003

xjas wrote:

Note: this system is running 98SE with KernelEX installed (the K6-2 doesn't have it.) Not sure that makes any difference. This is also a less "pristine" system meaning I've had it for a while & have swapped around graphics & sound cards/drivers a bit.
When saving a report I get a crash at Section 103 [menus]. This happens every time, including on clean boot.

I installed KernelEx v0.36a and it's OK on my system, which version do you have installed?

I was planning to use the remote debugger to look into this, but as I don't see the issue can't, do you have MSVC 6 and know how to use it please?

red-ray wrote:

bakemono wrote:

FWIW, SIV32L and SIV32X (either one) crashes on my Win2000 socket AM2+ system when it gets to "UEFI info" if I try to save a report.

Thank you for checking out SIV and I am rather surprised [uefi-info] is there at all on a Windows 2000 system. What happens if you do Menu->File->Save->Save Selected and deselect [uefi-info] before pressing [ Start ] ?

Please will you also do Menu->File->Save->Save System ROM and post the .ROM file in addition to the .TXT + .DMI files? My guess is that SIV is crashing while scanning the contents of the System ROM.

I had a crash even with uefi-info unchecked (and this time it caused the system to freeze completely, I had to reboot and do CHKDSK and the whole nine yards), so I also unchecked bios-make and bios-option as one of those appeared to be the culprit. That worked.

This system uses Win2000 extended kernel, unofficial video drivers, and a CPU that is not recognized by the BIOS. Plenty of unusual stuff to see in here, no doubt.