I was looking at detection of PS/2 mouse.. It's very simple, but.. that method is causing a reset on my system.. any check for legacy HW that I add is causing an issue somewhere...
The check for not listing removed hardware that I implemented a while ago didn't work well and it was causing skipping of present cards too. So I removed it and currently don't know how to skip this properly.

I downloaded latest version (6.1.0), and retested a FIC board fror CPU sped issue.
I have also dug yet another old 486 board (Shuttle HOT-409) and ran some tests with it.

Results are:
FIC 486-GVT-2
- Still have an issue displaying incorrect CPU speed on the screen (report is however OK.)
SHUTTLE HOT-409
- This is an early VL board with 30-pin simms, LIF socket for CPU, and no voltage regulator for fastter CPUs. It has installed latest unofficial BIOS with support for >504MB drives (I have used a Transcend TS2GIFD25 – a 2GB IDE SSD during tests). He board doesn’t like more than 1 VLB card being used. If used with 2 cards, system refuses to load OS or crashes. Loose timings and adding waitstates doesn’t help, and I have used VLB graphic card and ISA IDE/IO card for tests with exception of no.9 (vl combo card) and no.10 (2 vl cards, crashed soon after generating the raport). It came to my mind (after the tests ☹ ) that maybe slowing the clock to 25 MHz would help? The board doesn’t also like 40 MHz bus – UMC CPU crashed shortly after report was made, and Cyrix DX40 refused to even load DOS, so it was run at 33 MHZ with no problem.
- Generally HWiNFO works ok. And recognizes correctly hardware being used, with the exception of 100MHz and 133MHz chips, in which cases CPU frequency is displayed and reported incorrectly. A 100 MHz CPU was able do display&report correctly if Auto Confi function in BIOS was Enabled, but alas, system was not stable then ☹ .
I guess, that BIOS couldn’t recognize that fast CPU and set incorrect timings.
With timings set manually system works stable, but CPU speed is reported wrong. (CHKCPU however detests it correctly).

Logfiles are attached.
Regards

phantom_pl wrote on 2020-04-29, 09:56:

I downloaded latest version (6.1.0), and retested a FIC board fror CPU sped issue. I have also dug yet another old 486 board (Sh […]
Show full quote

I downloaded latest version (6.1.0), and retested a FIC board fror CPU sped issue.
I have also dug yet another old 486 board (Shuttle HOT-409) and ran some tests with it.

Results are:
FIC 486-GVT-2
- Still have an issue displaying incorrect CPU speed on the screen (report is however OK.)
SHUTTLE HOT-409
- This is an early VL board with 30-pin simms, LIF socket for CPU, and no voltage regulator for fastter CPUs. It has installed latest unofficial BIOS with support for >504MB drives (I have used a Transcend TS2GIFD25 – a 2GB IDE SSD during tests). He board doesn’t like more than 1 VLB card being used. If used with 2 cards, system refuses to load OS or crashes. Loose timings and adding waitstates doesn’t help, and I have used VLB graphic card and ISA IDE/IO card for tests with exception of no.9 (vl combo card) and no.10 (2 vl cards, crashed soon after generating the raport). It came to my mind (after the tests ☹ ) that maybe slowing the clock to 25 MHz would help? The board doesn’t also like 40 MHz bus – UMC CPU crashed shortly after report was made, and Cyrix DX40 refused to even load DOS, so it was run at 33 MHZ with no problem.
- Generally HWiNFO works ok. And recognizes correctly hardware being used, with the exception of 100MHz and 133MHz chips, in which cases CPU frequency is displayed and reported incorrectly. A 100 MHz CPU was able do display&report correctly if Auto Confi function in BIOS was Enabled, but alas, system was not stable then ☹ .
I guess, that BIOS couldn’t recognize that fast CPU and set incorrect timings.
With timings set manually system works stable, but CPU speed is reported wrong. (CHKCPU however detests it correctly).

Logfiles are attached.
Regards

Thanks for the extensive report !
I see that there are a few minor things I can improve, but I still have no idea what's causing the reduced clock speeds for >100 MHz reported. What happens if you close the Mainboard info screen and open again a few times? Will it still show the low clock?
Also a question regarding #10: this one had an ALG2228 card, what's its exact name, some SPEA ?

Mumak wrote on 2020-04-29, 11:48:

What happens if you close the Mainboard info screen and open again a few times? Will it still show the low clock?

Still no changes. Every time the clock shown is too low

Also a question regarding #10: this one had an ALG2228 card, what's its exact name, some SPEA ?

Exactly, SPEA V7 Vega Plus

phantom_pl wrote on 2020-04-29, 12:57:

Still no changes. Every time the clock shown is too low

This is totally weird.
HWiNFO measures the execution time of a certain instruction set, so if it reports wrong there can be 2 reasons only: either the clock has really changed or something messed-up with the timers.
And if you exit HWiNFO right after it reporting low clock and launch CHECKIT does that report correct clock?

phantom_pl wrote on 2020-04-29, 12:57:

Exactly, SPEA V7 Vega Plus

Thanks, will add reporting of the card.

I didn't know about this thread until today.. I will submit a report from the Win32 version using my retro daily driver (a cookie cutter Pentium MMX build) later today. I will also try the DOS version. Not sure if that actually crates a debug report but if it does I will try that as well.

appiah4 wrote on 2020-04-29, 13:18:

I didn't know about this thread until today.. I will submit a report from the Win32 version using my retro daily driver (a cookie cutter Pentium MMX build) later today. I will also try the DOS version. Not sure if that actually crates a debug report but if it does I will try that as well.

Thanks in advance. The DOS version doesn't create a debug output.

Mumak wrote on 2020-04-29, 13:03:

This is totally weird.
HWiNFO measures the execution time of a certain instruction set, so if it reports wrong there can be 2 reasons only: either the clock has really changed or something messed-up with the timers.
And if you exit HWiNFO right after it reporting low clock and launch CHECKIT does that report correct clock?

I agree, that's weird, and most weird is correct clock in the report, and too low on the screen.
CHECKIT - ??? I have never used this tool, unless you mean CHKCPU .
I use CHKCPU 1.25 as a reference tool , and it shows correct clock
I took a picture of screen with CHKCPU running on 486dx4100, but I don't remember whether it is FIC or Shuttle board (but it looks the same)

Sorry, I meant CHKCPU 😀 Would be interesting to know what it reports right after HWiNFO reporting low clock. Just to prove whether HWiNFO does perhaps mess-up something with the clock.

Mumak wrote on 2020-04-29, 13:39:

Sorry, I meant CHKCPU 😀 Would be interesting to know what it reports right after HWiNFO reporting low clock. Just to prove whether HWiNFO does perhaps mess-up something with the clock.

I have already checked it, and I'm sorry, CHKCPU is always correct 🙁

phantom_pl wrote on 2020-04-29, 14:19:

Mumak wrote on 2020-04-29, 13:39:

Sorry, I meant CHKCPU 😀 Would be interesting to know what it reports right after HWiNFO reporting low clock. Just to prove whether HWiNFO does perhaps mess-up something with the clock.

I have already checked it, and I'm sorry, CHKCPU is always correct 🙁

Crazy... looks like nothing is impossible in this vintage world 😁
Scratching my head.... is the clock wrong also if you disable Reset Methods 1-3 ?

Scratching my head.... is the clock wrong also if you disable Reset Methods 1-3 ?

Changing CPU detection methods doesn't correct anything

Mumak wrote on 2020-04-29, 15:02:

Crazy... looks like nothing is impossible in this vintage world 😁

Nothing, but one....
A system capable of perfectly running all DOS software, with proper speed, and ideal sound 😁

phantom_pl wrote on 2020-04-29, 15:44:

Mumak wrote on 2020-04-29, 15:02:

Crazy... looks like nothing is impossible in this vintage world 😁

Nothing, but one....
A system capable of perfectly running all DOS software, with proper speed, and ideal sound 😁

Please try with this build if it's any different:

Test windows opens on Monday.
It will be first on a "to do" list 😉

Tested DRx2-66 at 66 MHz using the ALi PANDA board in Win95c. report and debug attached.

The CPU name looks wrong in the heading field. Shows up as: Cyrix Cx486DRxý-101.

Report file from my Cyrix MII, only thing that is obviously wrong is the sound card, it has a ESS 1869 on board with drivers loaded and working but isn't detected.
A few other things I notice is that is says max level 2 cache is 512k but shows the 1MB the motherboard contains. The CPU voltage shows as 3.3 v on the CPU tab, but the correct 2.9v on the summary and mainboard tabs, not sure if I'm not just reading this wrong though and that is is I/O voltage shown on that tab?

BSA Starfire wrote on 2020-05-03, 13:17:

Report file from my Cyrix MII, only thing that is obviously wrong is the sound card, it has a ESS 1869 on board with drivers loaded and working but isn't detected.
A few other things I notice is that is says max level 2 cache is 512k but shows the 1MB the motherboard contains. The CPU voltage shows as 3.3 v on the CPU tab, but the correct 2.9v on the summary and mainboard tabs, not sure if I'm not just reading this wrong though and that is is I/O voltage shown on that tab?

Thanks for the report. The max cache size (512K) and 3.3 V come from BIOS, which is responsible for providing accurate information, but it's often not the case.
Could you please try this build if it recognizes the ESS card: www.hwinfo.com/beta/hwi32_625_4146.zip

Yep, that's fixed it 😀

In my case nothing is fixed 🙁
Both boards (FIC 486-GVT-2 and Shuttle HOT-409) report and display wrong CPU clock (FIC is almost there in the report, but not on the display).
Report + HWiNFO printscreens + CHKCPU before&after screens are attached.