I've managed to somewhat attenuate the VRM issue on my P133 system, I have opened the case and put a small fan blowing directly on the offending VRM, it used to run a bit warm with a non-MMX Pentium but now it's relatively cold to the touch with a bit of airflow towards it.

So far the system has fired up just fine, report will follow once it has finished transferring SIV (long story short, since the P133 no longer has a NIC in it, I have to transfer files through a null-modem cable from the Omnibook to the P133, it's not a elegant solution but at least it works)

Well that didn't work unfortunately, SIV straight up refuses to launch and instead throws a illegal operation error 😒 it says something among the lines that a fault happened in MSVCRT.DLL (I tried installing the VCredist thing five times to no avail, even tried copying one from a known good system...)

Running SIV on the Omnibook's Win98 install also results in the same problem... So something must be going on with this version of SIV...

Here:

Vynix wrote:

So something must be going on with this version of SIV...

Thank you for checking out the new SIV, please can you confirm that it's SIV32L 5.43 Beta-00 what you are trying to run?

I have just successfully run this on one of my Windows 9x systems so it must be an issue unique to your system. I can see it's crashing within MSVCRT.DLL, but this does not give me any clue as all as to what SIV might be doing and as SIV is 489,970 lines in size I can't sensibly guess. When this happens there are debugging options which can be used to give an indication of what SIV is doing which are specified in my initial post.

red-ray wrote:

[…]
Show full quote

1. Generating logs if there is an issue with SIV start-up.
2. If SIV does not start from a command window do SIV32L -DBGINI -DBGTIM > SIV_DBGOUT.log | more which will trace the SIV start-up to SIV_DBGOUT.log which I need to be able to figure out the issue
3. If there is an issue with [PCI Bus] do SIV32L -DBGINI -DBGTIM -DBGPCI > SIV_DBGOUT.log | more
4. If there is an issue with the GPU Infomation do SIV32L -DBGGPU > SIV_DBGOUT.log | more
5. If there is an issue with Disk SMART Monitoring on NT do SIV32L -DBGSDM > SIV_DBGOUT.log | more

Please will you do SIV32L -DBGINI -DBGTIM -DBGGPU > SIV_DBGOUT.log | more and post SIV_DBGOUT.log.

It was the S3 beta (Trio01), IIRC, attempting to create a debug log with it resulted in a blank "DBGOUT" file.

Do you think that there could be a driver that would interfere with SIV and cause it to crash? Because just before running it on the P133, I had to reinstall the sound card's driver (said sound card would be a CT4180 Vibra16C).

Either way if I'll try to get a debug log, well, get another one because here is the kicker: I had created one (with the command that you provided) but forgot to attach it to my previous message. Sorry for that little oversight.

Vynix wrote:

Do you think that there could be a driver that would interfere with SIV and cause it to crash?

Without a SIV_DBGOUT.log file it's impossible to sensibly comment as to what it might be. If it's down to a driver you installed it should also affect Beta-00, does it?

I just tried SIV32L 5.43 Trio-01 on a W9x system with an S3 Trio GPU and it ran OK.

I have just uploaded SIV32L 5.43 Beta-01 to SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + SiS + Matrox + XGI + old ATI + NVidia GPUs - Testing Help, how does it do?

Hang on, I'll go and get the debug log that I created yesterday, I'll try beta 01 and see how it goes as well.

Included are the two debug files that SIV made when I ran it using the command you supplied:

Unfortunately, SIV BETA01 didn't work as well, I'm still getting the same error.

Vynix wrote:

Unfortunately, SIV BETA01 didn't work as well, I'm still getting the same error.

Thank you, it looks like it's going wonky trying to figure out what the CPU is.

After pondering this I suspect I could use a CPUZ save file to run SIV in test mode and recreate the issue here, please will you generate and attach it?

Sure thing, here goes:

I tried in failsafe mode to exclude a faulty driver causing issues (given that Windows won't load 3rd party drivers in failsafe), but still no dice 🙄

Vynix wrote:

Sure thing, here goes:

Thank you and my suspicion was correct 😁, I instantly got the crash which was down to a combination of an oversight in the Intel 430VX chipset support and some setup code not being defensive.

I have just attached SIV 5.43 Beta-02 to SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + SiS + Matrox + XGI + old ATI + NVidia GPUs - Testing Help which should resolve the issue, please let me know for it does?

Oh good lorry I'm glad it wasn't a driver that caused this because I seriously was considering reinstalling W98 on the P133 🤣

Edit: SIV is being loaded up on the P133, null-modem file transfer sure is slow even at 115200 bauds.

Here goes:

Here are the logs:

Vynix wrote:

Here are the logs:

Thank you for running this and all the other tests. I am relieved SIV is finally working as it should and checking [GPU Info] all looks as I would expect.

Do you know of any other utilities that report the S3 Trio64V2/DX clock speeds and if so does SIV report the same values please?

SIV 5.43 Beta-02 also has improved clock reporting for old ATI Rage (pre Rage 128) GPUs and I next plan to add S3 Savage 4 support.

I should look at that soon enough I might have another PC with PCI slots that is a wee bit newer than the P133... Because for some reason Aida doesn't report the clockspeeds on the P133.

As I said there is a possibility that I get one of these "Magma PCMCIA to PCI chassis", and stick the Trio64V2-DX I have in it, so I could try to get the clockspeed readings from the Omnibook.

Vynix wrote:

there is a possibility that I get one of these "Magma PCMCIA to PCI chassis"

Given the price would it not be cheaper to get a P-III desktop system?

I have SIV reporting my S3 Savage4 Pro-M, but there are 86C394/5/6/7/8 chips I should test SIV on and I can't even find the chip ID of the 86C398.

They all have the same PCI Device ID so you need to poke around the GPU SR registers to tell which chip it is.

The price of these is sure steep, but it's not really a problem for me, in fact, I've been saving up to one for a later project.

As for getting another P3 desktop system, I've considered this as a backup plan, but right now I'm stuck with laptop systems (exception of the P133 which is my only desktop system).