red-ray wrote on 2025-04-29, 00:50:

BetaC wrote on 2025-04-28, 21:24:

No, it is a SX731. As for what SIV says, this is what I got out of Windows 95. Would an NT install help in this instance?

Thank you and Windows NT would not really help, I expect the stepping would be D0 rather than 0, but most DX CPUs that don't have CPUID return D0 on NT, SX return C0 and the 486DX-25 SX308 is the only CPU for which I have seen B0, see Early 80486 in https://www.geoffchappell.com/studies/windows … pu/precpuid.htm, I wonder which i486 returns A0. This is the first time I am aware of SIV correctly reporting a DX2 that does not support the CPUID instruction.

The SX731 and SX759 both have a CPUID of 0433 so are effectively identical. I have updated SIV so Beta-06 will report both SX731 and SX759 have a blue heatsink.

Awesome, happy to have helped.

I have just released SIV V5.81 Beta-06, the changes are as below.

5.81 Beta-06 managed to work out that a SX308, which does not have CPUID, is an Intel 486 DX (P4) [B] Model 00 Stepping 1 and even that it is a SX308! SIV could deduce this as Windows NT returned FFB0 and for CPUID 0404 and later it's FFD0, but I don't know what CPUID 0402 will return. If you have a SX328 or SX329 running NT please will you post what Beta-06 reports.

With the SX308 installed H3nrik V!’s system is non too stable, but is with several other CPUs we have tried. I am wondering, do SX308s in general have issues running NT? If you know please enlighten me.

I am also trying to find out what the CPUID is for SX249 + SX250 + SX493 + IBM PN 63G9145 + SX412, AFAIK none have CPUID support. Is there a DOS program that can get the CPUID for these CPUs by doing a reset and getting the CPUID from EDX on the restart?

I have just released SIV V5.81 Beta-07, the main change since Beta-06 is the addition of [System Summary] which provides an overview of the Windows version, CPUs including cache topology, memory, disk drives, GPUs and motherboard. A Left/Click on the Windows logo will popup this panel and it's also available as View Summary via the [Copy|▼] dropdown or Right/Click for pre Vista systems.

Note: As ES_CENTER is not supported for simple edit controls on NT V4.00 they are left justified.

Is it normal that HPET reads zero on Win 2000? I believe there is no OS support for HPET until XP SP3, so maybe this is expected behavior. (under hardware/cpu detail/cpu timers)

bakemono wrote on 2025-05-11, 16:45:

Is it normal that HPET reads zero on Win 2000?

Thank you for the report and I guess so, I don't have a system that's running Windows 2000 with a HPET though, but I just found a SIV save file from a Intel Core 2 Q9550 (Yorkfield) system running NT V3.51 and again the HPET is stopped. The only use SIV makes of the HPET if to report it on [CPU Timers] so it's possible to see if there any dilation, given this it not running is no big deal. I seem to recall CPUZ used to use/start the HEPT and this caused issues when Windows XP started using it, but that was 20 years ago so an none too sure. I wonder if CPUZ will start it on pre Windows XP systems.

I also noted that the QPC is the same speed as the TSC, so is the TSC, with later versions on Windows this is not the case. With Windows 2000 this is not an issues as SIV uses the Media clock for CPU speed measurements, but from XP-64/2003 it uses the QPC so it being the TSC is sub-optimal.

Update: I remembered I had Windows 2000 Advanced Server on my Supermicro X7SBL Intel Xeon X3320 QC (Yorkfield) so I booted it and the HPET was stopped, but when I ran HWiNFO32 it started it and it remained running even after I had exited HWiNFO! Out of interest how many CPUs show up? I installed AS as I think with professional I only got 2 and 3.25GB of memory, I now get all 4 and 7.75GB .

red-ray wrote on 2025-05-11, 18:10:

Update: I remembered I had Windows 2000 Advanced Server on my Supermicro X7SBL Intel Xeon X3320 QC (Yorkfield) so I booted it and the HPET was stopped, but when I ran HWiNFO32 it started it and it remained running even after I had exited HWiNFO! Out of interest how many CPUs show up? I installed AS as I think with professional I only got 2 and 3.25GB of memory, I now get all 4 and 7.75GB .

Running HWiNFO32 or CPUZ didn't seem to have any effect on the HPET for me.
With the Extended Kernel, all 6 cores are recognized (I've been told it can go up to 32), and PAE mode is also available. Though I'll note that some programs report 1 core, 6 threads. I suspect this is based on the result of GetLogicalProcessorInformation, and the fact that Win2K Pro is restricted to 2 "processors" but not threads. To be clear, Win2K in its original form can't boot on this board with the ACPI Multiprocessor HAL, so hacks are needed to get it up and running with more than a single core.

bakemono wrote on 2025-05-12, 14:56:

Running HWiNFO32 or CPUZ didn't seem to have any effect on the HPET for me.
With the Extended Kernel

Thank you for your reply, I also found CPUZ had no impact, which HWiNFO32 did you use? I used 8.25-5720, maybe the Extended Kernel has an effect on it.

Now I remember you! I recall issues caused by the Extended Kernel half supporting GetFirmwareEnvironmentVariable(), do I recall correctly?

I wonder why Cinebench came up 1 + 6, please will you do Menu->File->Save Local and post/pm/e-mail be the two save files so I can try and figure it out, maybe I have one, what was your system called, I can't find one for ZEN2000. GetLogicalProcessorInformation() came in with XP SP3.

Update: Found it, [numa] is not in the old Save File, I have just tweaked SIV so that if GetLogicalProcessorInformation() is in kernel32 the NUMA information will get reported, please can you use depends and check if it is or ZIP and post kernel32.dll so I can?

red-ray wrote on 2025-05-12, 15:41:

Thank you for your reply, I also found CPUZ had no impact, which HWiNFO32 did you use? I used 8.25-5720, maybe the Extended Kernel has an effect on it.

It was 8.24-5700

Now I remember you! I recall issues caused by the Extended Kernel half supporting GetFirmwareEnvironmentVariable(), do I recall correctly?

Correct. I'm not sure if that issue was ever fixed or not. This is a different system though in any case.

Update: Found it, [numa] is not in the old Save File, I have just tweaked SIV so that if GetLogicalProcessorInformation() is in kernel32 the NUMA information will get reported, please can you use depends and check if it is or ZIP and post kernel32.dll so I can?

It's in there. I'll attach a list of all kernel32 exports .

BTW, I'd be curious if you know what is going wrong with ACPI methods. I think it's related to the acpi.sys from Windows XP Integral Edition.

bakemono wrote on 2025-05-13, 15:22:

I'd be curious if you know what is going wrong with ACPI methods. I think it's related to the acpi.sys from Windows XP Integral Edition.

Thank you for the save files. I noted SIV half thought your Ryzen 5 was a Ryzen 3, I have just adjusted the code to fix this.

The GetLogicalProcessorInformation() was not in the save file, I plan to release SIV V5.81 tomorrow, once I have released it do Menu->Hardware->NUMA Nodes and post the panel.

As for ACPI the OBJDATA and NSOBJ layouts are different between V5.00 and V5.01 and as your system is build 2195 SIVDriver is using the V5.00 ones and I suspect your system need it to use the V5.01 ones, if I could get SIVDriver to use the V5.01 code when the acpi.sys from Windows XP was being used all the wonky method names would probably go away. In truth I am half surprised you didn't get a BSOD! What we need is a dump of the ACPI PDO Extension from your system and to compare that with the ones from standard W2K + WXP systems to see if there is a way SIVDriver can know it should use the V5.01 code. Do SIV32X -WDF then Menu->System->ACPI + PnP->ACPI Extension, then press [Copy], pate the text into a file and post it. I will compare it to standard W2K + WXP ones and may be able to safely adjust SIVDriver and give you a test SIVX32.sys to try out.

SIV Version 5.81 released on 14-May-2025 - I just released it, the main changes as as below:

1. Added [System Summary] which provides an overview of the Windows version, CPUs including cache topology, memory, disk drives, GPUs and motherboard. A Left/Click on the Windows logo will popup this panel and it's also available as View Summary via the [Copy|▼] dropdown.
2. Added AMD Ryzen (Hawk Point) + (Rembrandt) per CPU Core + L3 Cache temperature reporting.
3. Improved support for i386 + i486 CPUs that don't have CPUID, especially so for Windows 9x and FPU detection.
4. Added the [CPU CR0] panel and corrected the MB/Second when Seconds is > 4.095 on [Memory Speeds]
5. Added the [CPU i486 S-Spec] panel, S-Specs that have been added to CPUDEVS.txt are reported in the CPUs in CPUDEVS colour.
6. Added the [CPU S-Spec Setup] panel which allows the user to setup the CPU S-Specs and specifies which save file filename suffixes to apply.
7. Updated [GPU Temperatures] to use the current temperature selection and added °O Onkian + °S SImyon support.
8. Added Menu->Help->Conversions->CPU Voltage R1 R2 which shows the R1 and R2 values needed to set non standard voltages on my PGA-DX4/2.
9. Resolved issue with the [Hardware Status] panel crashing on some systems with an odd number of DIMMs.
10. Resolved issue with tooltips not working when SIV32X was run on Windows 2000 and added SIV32N, a Unicode SIV for Windows NT V4.00.
11. Added ASRock 990FX Extreme3 + ASUS CUBX + Chaintech 4SOM2.1 (M102) motherboard support. Fixed MSI MS-6135 -12 and -5 voltage reporting.
12. Added Windows 11 25H2 Build 27813 support. Enabled operation of [ACPI Eval], etc..

SIV 5.81 can be downloaded via https://workupload.com/file/VuuN4yRsEgY

red-ray wrote on 2025-05-02, 19:49:

I don't know what CPUID 0402 will return. If you have a SX328 or SX329 running NT please will you post what Beta-06 reports.

Mystery solved, my SX329 just arrived and it's FFC0, my system is also stable @ 33 MHz and will even run @ 40 MHz, but Windows NT V4.00 won't boot when it's @ 40 MHz.

After some minor adjustments SIV reported as below.

red-ray wrote on 2025-05-02, 19:49:

I am also trying to find out what the CPUID is for SX249 + SX250 + SX493 + IBM PN 63G9145 + SX412, AFAIK none have CPUID support.

I am sill wondering about the unknown CPUIDs and H3nrik V! is wondering if all SX308 are unstable or just his so reports about any of these CPUs would be good.

lolo799 wrote on 2025-05-27, 20:00:

I did it earlier and it crashed at the sm-bus part, I'll retry another time and post it in your SIV thread.

Thank you, did SIV crash or did the system crash/reboot? If the latter there is probably a strange SMBus device, please do: SIV32X -DBGSMB -SAVE=[smb-bus]=SIV_DBGOUT.txt > SIV_DBGOUT.log | more and post the two SIV_DBGOUT files.

I suspect the last line of SIV_DBGOUT.log will be the strange/problem device, you could do Menu->Tools->Configure->SIV SMBus Setup and add an exclusion for it then try as above again in case there are multiple strange devices.

For the save do Menu->File->Save->Save Selected, deselect [smb-bus] then press [Start].

red-ray wrote on 2025-05-27, 20:10:

Thank you, did SIV crash or did the system crash/reboot? If the latter there is probably a strange SMBus device, please do: SIV […]
Show full quote

lolo799 wrote on 2025-05-27, 20:00:

I did it earlier and it crashed at the sm-bus part, I'll retry another time and post it in your SIV thread.

Thank you, did SIV crash or did the system crash/reboot? If the latter there is probably a strange SMBus device, please do: SIV32X -DBGSMB -SAVE=[smb-bus]=SIV_DBGOUT.txt > SIV_DBGOUT.log | more and post the two SIV_DBGOUT files.

I suspect the last line of SIV_DBGOUT.log will be the strange/problem device, you could do Menu->Tools->Configure->SIV SMBus Setup and add an exclusion for it then try as above again in case there are multiple strange devices.

For the save do Menu->File->Save->Save Selected, deselect [smb-bus] then press [Start].

The system froze completely.
I had to deselect pmb-bus too, otherwise the same thing happened.

lolo799 wrote on 2025-05-28, 17:59:

The system froze completely.

Thank you for the save files and apologies for SIV freezing the system. Below you can see SIV V5.82 Beta-02 should be fixed and report Intel Pentium II Xeon (Drake).

Looking at the -DBGSMB output it's not that clear what the root cause it.

1PWMI PWM    Slave 0_70  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 64   Size 32
2PWMI PWM    Slave 0_60  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
3PWMI PWM    Slave 0_41  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 64   Size 32
4PWMI PWM    Slave 0_45  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
5PWMI PWM    Slave 0_20  Regime 3  Model 9A9A  CMD 3D  Numb 2    Status 0002  Time 226  Size 34
6PWMI PWM    Slave 0_78  Regime 2  Model 9A9A  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
7PWMI PWM    Slave 0_40  Regime 2  Model 9A9A  CMD 9A  Numb 0    Status 0004  Time 62   Size 32
8PWMI PWM    Slave 0_71  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
9PWMI PWM    Slave 0_72  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 64   Size 32
10PWMI PWM    Slave 0_73  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
11PWMI PWM    Slave 0_74  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 64   Size 32
12PWMI PWM    Slave 0_75  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
13PWMI PWM    Slave 0_76  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32
14PWMI PWM    Slave 0_77  Regime 2  Model 0000  CMD 9A  Numb 0    Status 0004  Time 63   Size 32

My best guess is that [ 0_20 ] is the root cause as from the Time 226 there seems to be a device there and Model 9A9A also being reported for [ 0_78 ] + [ 0_40 ] is strange, further there is nothing after [ 0_77 ] when it should be trying to read the DIMM SPD + Xeon PIROM information.

1. Use Menu->Tools->Configure->SIV SMBus Setup to add an exclusion for Slave [ 0 _ 20 ]
2. Exit SIV and reboot the system
3. Do SIV32X -DBGSMB -SAVE=[smb-bus]=SMBus.txt > SMBus.log | more and post the two SMBus files.
4. Do SIV32X -DBGSMB -SAVE=[pmb-bus]=PMBus.txt > PMBus.log | more and post the two PMBus files.
5. If the system is still running do SIV32X -DBGSMB > SIV_DBGOUT.log | more then press [ SPD ], is the SPD reported?
6. To check the Xeon PIROMs do Menu->Hardware->MSRs->Xeon PIRs.
7. I hope/suspect [Save Local] will now be OK.

Update: SIV V5.82 Beta-02 should automatically exclude [ 0_20 ], use it and skip #1

i create a file for you from my Voodoo setup uses a cor2duo on win98se. Siv really struggle with that setup 😀

red-ray wrote on 2019-08-22, 10:17:

SIV is designed for Windows NT so if you have both Windows NT/XP and 9X use the NT/XP system as more will be reported. Chances are that if something is zero or absent on 9x/Me the correct value will be reported on NT 3.51 and later.

Jackhead wrote on 2025-05-29, 14:25:

i create a file for you from my Voodoo setup uses a cor2duo on win98se. Siv really struggle with that setup 😀

Thank you, I specified SIV will never do as well on 9x as on NT so this is expected.

Which ASUS motherboard is it? SIV could use the 8276 to detect this and the SIO reporting should get better.

its a Asus P5K-WS X55 with an E8600

Jackhead wrote on 2025-05-29, 14:58:

its a Asus P5K-WS X55

Thank you, SIV V5.82 Beta-02 should be less bad. I may further improve things, I released it for the new Windows 11 25H2 Build 27868 support

red-ray wrote on 2025-05-28, 20:56:

Thank you for the save files and apologies for SIV freezing the system. Below you can see SIV V5.82 Beta-02 should be fixed and report Intel Pentium II Xeon (Drake).

The system still froze at the pmb-bus part when doing a file->save->local.
The MSRs menu didn't show Xeons in the list.
Not sure why the 2nd cpu disappeared, i had to completely turn off the system and restart it for it to show up again.

lolo799 wrote on 2025-05-30, 13:37:

The MSRs menu didn't show Xeons in the list.
Not sure why the 2nd cpu disappeared, i had to completely turn off the system and restart it for it to show up again.

Thank you for running the tests. There will only be an entry in the MSRs menu if SIV detects the Xeon PIRs.

What's happening is really strange as there should be a report in the .LOG files for the problem device, further SIV can't even read the DIMM SPD, do any programs report the DIMM SPD?

For some reason HKCU\Software\SIV SMBus Skip has been created, please delete it then restart SIV, see Menu->Machine->SIV Setup->SIV Setup Global.

I feel that the issue may be triggered by the PWM detection so I have built SIV32X V5.82 Beta-03 that does not do this and also adds several SMBus Exclusions, using it please do SIV32X -DBGSMB -DBGSPD > SIV_DBGOUT.log | more. I am hoping it finally read the SPD and Xeon PIRs.