@red-ray, Had to run around through some menus to find PCI Bus, as it seems the main window really doesn't fit on 800x600 resolution.
Though, when i clicked the PCI Bus, once again, the application freezed. (Just application itself)

It's clean install of Windows 95C, the OSR 2.5, latest windows 95.

GigAHerZ wrote:

@red-ray, Had to run around through some menus to find PCI Bus, as it seems the main window really doesn't fit on 800x600 resolution.
Though, when i clicked the PCI Bus, once again, the application freezed.

Thank you, but it should more or less fit, can you post a screen shot? Below is what I get on my NT 3.10 system which is 800 x 600 as I can't find a driver for it's Matrox Millennium G450 Dual Head PCI - Vendor ID 102B-0525.

I am wondering if it's scanning all 256 PCI buses and just taking a very long time as on start-up SIV should have read the [MCH]. From a command window do SIV32L.exe -DBGINI -DBGPCI > SIV_DBGOUT.log | more and post SIV_DBGOUT.log.

Once I see this I may be able to speed things up.

@red-ray,
Hope i understood the instructions correctly.

GigAHerZ wrote:

Hope i understood the instructions correctly.

Thank you, almost ideal and in truth SIV did better then I suspected it would, but I wonder why it's Family 1 😕. The width is down to the font and I suspect if you install the attached Tahoma font it may do better. You can use Menu->Tools->Font to change the font.

Looking in SIV_DBGOUT.log I can see SIV did try and read some PCI devices and expect if you wait long enough [PCI Bus] would complete, but it may take a very long time, I think over 9 hours 😲

Please will you run SIV again with the debug options and do Menu->File->Save->Save Selected, uncheck [pci-bus] + [pci-all] then press [ Start ] to do the save and post the save file + the new SIV_DBGOUT.log.

Once I get these then I will adjust SIV some more.

@red_ray, there you go!

GigAHerZ wrote:

@red_ray, there you go!

Thank you, it's now all starting to half make sense. Did changing the font to Tahoma make the panel fit?

Using the attached SIV 5.41 Skip-14 please redo the SIV32L.exe -DBGINI -DBGPCI > SIV_DBGOUT.log | more and post SIV_DBGOUT.log + save file, but this time just do Menu->File->Save Local.

With a bit of luck SIV should report the CPU type and speed, is it correct? For the speed it would be helpful if you posted the initial screen and Menu->Help->Conversions->CPU MHz.

I adjusted [PCI Bus] so it should be faster and also added a 30 second timeout. I need to check the new SIV_DBGOUT.log to see if it worked.

red-ray wrote:

Thank you, maybe NT 3.51 is different, please try SIV32L.exe -DBGINI -NOSIVRES > SIV_DBGOUT.log | more. Ideally use Beta-12 as I […]
Show full quote

feipoa wrote:

SIV_DBGOUT.log was not generated in the SIV folder. I'm using the original version you posted, not the beta for NT.

Thank you, maybe NT 3.51 is different, please try SIV32L.exe -DBGINI -NOSIVRES > SIV_DBGOUT.log | more. Ideally use Beta-12 as I added some extra debug traces.

Beta-11 was not ideal so I made some more changes for the attached Beta-12, please try it on W95C (It should say 95C). Again use SIV32L.exe -DBGINI > SIV_DBGOUT.log | more so I can see how the detection is done.

I tried to install NT 3.51 on one of my systems, but thus far the install fails. I suspect I need to use an older system. Do you have any other systems with NT 3.51 installed and if so how does SIV do?

EDIT: I think I have found the issue with NT 3.51. It seems to have issues loading images from SIVRES.dll so please try SIV32L.exe -DBGINI -NOSIVRES > SIV_DBGOUT.log | more on 3.51

Beta 12 still generates a Dr. Watson in NT 3.51, but I was able to run, SIV32L.exe -DBGINI -NOSIVRES > SIV_DBGOUT.log | more

See attached log file and screenshot.

After running this command I was unable to shutdown the system. I try to shutdown normally via the GUI (see image below), but after hitting OK, the system continues running as if I never asked it to shutdown.

Concerning Windows 95c, I no longer needed to run IDON.COM to use SIV32L, so whatever you adjusted worked.

CuPid wrote:

@feipoa […]
Show full quote

feipoa wrote:

For NexGen information, I suggest you contact the guy who compiled this website, https://www.memotech.franken.de/NexGen/ He goes […]
Show full quote

For NexGen information, I suggest you contact the guy who compiled this website, https://www.memotech.franken.de/NexGen/
He goes by mtx500 on the CPU-World forum.

I can try a pure 386 instruction set if you want to compile it for that. Might fix the issue, might not. Generally, it would be nice not to use IDON.COM as it can cause issues with other programs, but if some program doesn't work naturally, I load IDON.COM, then run the program.

Do you want me to try cpuz_w95 again on NexGen in Windows 95? And which version of CPU-Z is to be used for NT4 on a 486 or NexGen? Is that the same version that hopefully will work with NT 3.51?

@feipoa

Just to check if that changes anything with ICON.COM, I've recompiled with the "386 instruction set" setting (instead of "blend" by default).
http://download.cpuid.com/betas/cpuz_386.zip

Thanks !

Unfortunately it didn't help. I still need to run IDON.COM. Perhaps communicate with red-ray to discover how he was able to get SIV32L to not need IDON.COM loaded.

This is how it looks on Al-Jalima running NT 4.0 SP6a. I can actually run the Win98 version just fine, but this version manages to catch some aspects of my system that version can't, like my actual motherboard model and its chipset. Please consider letting us bench against some more powerful chips like the PII-450 and such, the most powerful chip I saw in the drop-down menu was a PMMX166 that I wiped the floor with, especially in FPU.

feipoa wrote:

Beta 12 still generates a Dr. Watson in NT 3.51, but I was able to run, SIV32L.exe -DBGINI -NOSIVRES > SIV_DBGOUT.log | more

After running this command I was unable to shutdown the system. I try to shutdown normally via the GUI (see image below), but after hitting OK, the system continues running as if I never asked it to shutdown.

Concerning Windows 95c, I no longer needed to run IDON.COM to use SIV32L, so whatever you adjusted worked.

Thank you for running the tests and I am quite pleased that the crash has gone which was basically caused by SIVRES.dll containing more images than NT 3.51 can handle. If you use SIV32L 5.41 Beta-15 (download/file.php?id=67253) then -NOSIVRES should not ne needed. Beta-15 should also report the CPU as NexGen and get the speed correct.

Using 5.41 Beta-15 on NT 3.51 please will you do SIV32L -DBGINI > SIV_DBGOUT.log | more, generate the two save files (Menu->File->Save Local) and either post or email them to me along with SIV_DBGOUT.log? The .DMI file is a dump of the BIOS data that's used to generate the [Machine] panel. Once I get these I suspect I will be able to improve the motherboard reporting.

If you do Menu->Machine->Shutdown what happens? Was SIV32L still active when you tried to do the shutdown? Can you post the Task Manager screenshot so I can see which programs are active. Post Menu->Edit->View STC Info so I can see what the SIV worker threads are doing.

I did not make any changes to address the IDON.COM issue which I think is down to the code in SHLWAPI.dll. Lets see what happens when you generate the SIV save file.

Please reply to SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + SiS + Matrox + XGI + old ATI + NVidia GPUs - Testing Help

Sedrosken wrote:

This is how it looks on Al-Jalima running NT 4.0 SP6a. I can actually run the Win98 version just fine, but this version manages to catch some aspects of my system that version can't

I suspect most of this is down to CPUZ 1.89.3 being newer than the Windows 98 version you have.

I noticed you have a 3Dfx Voodoo Banshee GPU and am hoping you will try my SIV utility and check if it reports the GPU Clocks. To make it easy I have attached SIV 5.41 Beta-14.

Please post the initial screen + Menu->Machine->GPU Info so I can check all is OK, do the CPU Clock speeds look correct to you? Ideally also do Menu->File->Save Local and post the two save files so I can check a few other things are OK as SIV does not get tested much on 3rd party NT4 systems these days.

Please reply to SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + SiS + Matrox + XGI + old ATI + NVidia GPUs - Testing Help

feiopa and Sedrosken : thanks for your feedback.
I tried to get rid of SHLWAPI but it seems quiite complicated.

GigAHerZ :

I fixed the 40 MHz on the DLC, and the PCI lock issue (at least, I tried to). That version should be OK w/o the ini file trick.
http://download.cpuid.com/betas/cpuz_w95.zip

Thanks !

CuPid wrote:

I tried to get rid of SHLWAPI but it seems quite complicated.

I suspect what CPUZ and SIV should do is Enable NexGen 486 Emulation, but figuring out what is needed will be tricky unless we can run a debugger on a NexGen CPU.

I have looked at what http://www.cpu-world.com/forum/viewtopic.php?t=31984 says and put the code from ftp://retronn.de/driver/CPU/nexgen/pci into SIV then looked at the decompile.

The mov esi,3B0B00Eh + mov ah,4Ch make sense, but the rest of the code does not. The emit 0x0F emit 0x55 is the start of the enable, but how long is it?

It would be interesting to single step through the code on a NexGen CPU and suspect after doing this it would be possible to add some code to CPUZ + SIV such that IDON.COM did not need to be run.

I have just realised the code is for 16-bit mode so will need to be adjusted for 32-bit mode.

What happens if IDON.com is run from a command window on NT 3.51?

@red-ray, @CuPid, here are the latest runs.

For cpu-z, i didn't need the ini file any more and the benchmark did run in around 10 seconds.

GigAHerZ wrote:

@red-ray, @CuPid, here are the latest runs.

For cpu-z, i didn't need the ini file any more and the benchmark did run in around 10 seconds.

Thanks ! 😎

GigAHerZ wrote:

@red-ray, @CuPid, here are the latest runs.

Thank you GigAHerZ, SIV did better, but as it got the family wrong it failed to detect it as a Cyrix CPU. Please try with the attached SIV32L 5.41 Beta-15

Please do SIV32L.exe -DBGINI -DBGPCI > SIV_DBGOUT.log | more and post SIV_DBGOUT.log + save file generated by doing Menu->File->Save Local.

I adjusted [PCI Bus] so it should be faster and also added a 30 second timeout. I need to check the new SIV_DBGOUT.log to see if it worked. You need to do the Save Local so the PCI Bus get's scanned.

Please reply to SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + SiS + Matrox + XGI + old ATI + NVidia GPUs - Testing Help

I ran it on my ComputerLab ST5500x thin client (VIA Eden-based.) A couple minor issues:

- the CPU is reported as a "VIA Eden 500MHz" running at 1GHz (5 x 200). I think this is a problem with the CPUID as it happens in everything, including SIV, Windows's device manager, and Linux's /proc/cpuinfo.

- this build of CPU-Z can't detect the RAM frequency and shows no info (blank page) under SPD. CPU-Z 1.86 x32 doesn't have this problem:

- this build takes much longer to start up than 1.86 and seems to spend a long time around 20-25% when it tries to read the SPD. Probably related to above.

Here's the benchmark:

Note that this machine runs at 1GHz, not 500MHz.

System report for the above machine:

Thanks xjas.

Yes your cpu does well run at 1GHz. The processor name string is not well programmed (most likely a BIOS issue).
I made some fixes, hoping that this solves most of the problems that you had.

http://download.cpuid.com/betas/cpuz_w95.zip

Just downloaded the last version above here. Intel 486DX2-66 (SX750) is recognized as an AMD chip. Biostar MB-8433UUD-A with Windows 95.