CoffeeOne wrote on 2020-02-12, 14:24:

Mumak wrote on 2020-02-12, 14:08:

Thanks for the feedback and all the test runs! Great to see that the EISA scan finally works as expected :) IIRC there were alwa […]
Show full quote

CoffeeOne wrote on 2020-02-12, 13:53:

I am back :) Looks already quite good. I will attach a screenshot, EISA devices are listed, will also attach the debug file. […]
Show full quote

I am back 😀
Looks already quite good.
I will attach a screenshot, EISA devices are listed, will also attach the debug file.

I checked again the benchmark: I select only "CPU" and let it run => then I get "NA". After pressing the button close, windows crashes, I have to reboot.

As mentioned before already, the dos version hwinfo555 hangs in showing memory information (maybe running out of memory?), but that's a minor point, I like the windows version more.

Thanks for the feedback and all the test runs! Great to see that the EISA scan finally works as expected 😀
IIRC there were always some issues with benchmarks on Win9x but I didn't find the cause yet. Will have another look at that.
As for the DOS version I'm no longer maintaining this, but perhaps some day I will get back to this...
Another thing that I was thinking about was adding the SB16 detection. This might work under NT systems, but for W9x I'd need a dump of the registry (HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet). Would you be willing to do this?

Yes

Thanks! Unfortunately it looks like the values I expected are not present in that node. Perhaps it's somewhere else under HKEY_LOCAL_MACHINE or not supported at all on Win9x.

Mumak wrote on 2020-02-12, 14:44:

CoffeeOne wrote on 2020-02-12, 14:24:

Mumak wrote on 2020-02-12, 14:08:

Thanks for the feedback and all the test runs! Great to see that the EISA scan finally works as expected :) IIRC there were alwa […]
Show full quote

Thanks for the feedback and all the test runs! Great to see that the EISA scan finally works as expected 😀
IIRC there were always some issues with benchmarks on Win9x but I didn't find the cause yet. Will have another look at that.
As for the DOS version I'm no longer maintaining this, but perhaps some day I will get back to this...
Another thing that I was thinking about was adding the SB16 detection. This might work under NT systems, but for W9x I'd need a dump of the registry (HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet). Would you be willing to do this?

Yes

Thanks! Unfortunately it looks like the values I expected are not present in that node. Perhaps it's somewhere else under HKEY_LOCAL_MACHINE or not supported at all on Win9x.

Full reg

CoffeeOne wrote on 2020-02-12, 14:54:

Mumak wrote on 2020-02-12, 14:44:

CoffeeOne wrote on 2020-02-12, 14:24:

Yes

Thanks! Unfortunately it looks like the values I expected are not present in that node. Perhaps it's somewhere else under HKEY_LOCAL_MACHINE or not supported at all on Win9x.

Full reg

Thanks! Unfortunately no luck either. It looks like the W9x driver doesn't provide some of the interesting data in registry.

Slashzero wrote on 2020-02-11, 03:40:

Sorry, posting this from Win95 on the 486. HWiNFO will not run on Win95... I will post some screenshots in a while. […]
Show full quote

Sorry, posting this from Win95 on the 486. HWiNFO will not run on Win95... I will post some screenshots in a while.

It installed fine, but complains about a missing DLL on startup.
53A45DFA-BDAB-44BC-9678-8498CF7A5341.jpeg

I realize I just assumed it would work with Win95 since you asked for feedback from a 486. Is that not supported?

Can you try to install some newer/latest version of IE if that will fix this message?
Apparently a lot of required DLLs get installed with it and that would make the life much easier 😀
I checked a Win95B Build 1212 installation and that seems to contain the required DLLs. What's your version?

Mumak wrote on 2020-02-12, 15:35:

Can you try to install some newer/latest version of IE if that will fix this message?
Apparently a lot of required DLLs get installed with it and that would make the life much easier 😀
I checked a Win95B Build 1212 installation and that seems to contain the required DLLs. What's your version?

I am using WIN95B (OSR). Yesterday I installed IE4. I will try HWiNFO again tonight if I have a chance. No luck. Even though I installed IE4, it still won't run.

Can you look into Intel Pentium D 965 EE and Xeon 5080 dempsey issues:

When overclocking they both report wrong cpu speed bus clock when overclocked with custom multiplier in bios. In my case, the correct values should be 12x multiplier, fsb is 375 which drives a 4500Mhz cpu speed and the dram should be 600mhz. Instead it reports 12x 321 fsb and a speed of 3857Mhz with a dram speed of 514Mhz.

Xeon 5060 with an default multiplier of 12x in opposite of the 14x default multiplier for the cpu's above, does not have this issue. Speedstep maybe being the cause for issue in HWiNFO as it is assuming speedstep even when speedstep is not actually used on the xeon 5080 and the D965 EE?

DBG file http://dottoss.com/files/HWiNFO64.DBG

dottoss wrote on 2020-02-15, 11:46:

Can you look into Intel Pentium D 965 EE and Xeon 5080 dempsey issues: […]
Show full quote

Can you look into Intel Pentium D 965 EE and Xeon 5080 dempsey issues:

When overclocking they both report wrong cpu speed bus clock when overclocked with custom multiplier in bios. In my case, the correct values should be 12x multiplier, fsb is 375 which drives a 4500Mhz cpu speed and the dram should be 600mhz. Instead it reports 12x 321 fsb and a speed of 3857Mhz with a dram speed of 514Mhz.

Xeon 5060 with an default multiplier of 12x in opposite of the 14x default multiplier for the cpu's above, does not have this issue. Speedstep maybe being the cause for issue in HWiNFO as it is assuming speedstep even when speedstep is not actually used on the xeon 5080 and the D965 EE?

DBG file http://dottoss.com/files/HWiNFO64.DBG

Yes, both C1E or EIST (SpeedStep) are affecting this and it's intentional behavior. When those are not active, HWiNFO will report the expected clock.
According to the dump you provided, EIST is enabled.
What clock does CPU-Z report?

CPU-Z shows the correct speed of 4500Mhz and correct FSB of 375 and also correct DRAM speed of 600 (1200mhz). The MB, CPU or whatever it is might report EIST enabled but it is in effect not functional as the lowest multiplier the chip can do is 12x which has been manually already set in bios. It can't go lower than that but it seems hwinfo is calculating it either way by lowering the bus calculation. I'll get you a screenshot of CPU-Z.

Here you go

Actually i'm not even sure EIST is working properly with these cpus even tho Intel specs it on their website as i've never seen it regulate fsb even when not overclocked, just the multipliers between 14-12x, ie C1E.

Source: https://www.anandtech.com/show/1975

EDIT:

Pentium D 965 according to intel spec does not support EIST
https://ark.intel.com/content/www/us/en/ark/p … 66-mhz-fsb.html

But XEON 5080 does, which is curious as they are basically the same chip.
https://ark.intel.com/content/www/us/en/ark/p … 66-mhz-fsb.html

Eitherway, I get the same incorrect speed reports regardless if I use a 965EE or the 5080.

EIST doesn't affect FSB, the issue here is different. HWiNFO properly calculates the actual clock (4500Mhz), but then thinks this rate is based on max clock (x14) and needs to be adjusted due to EIST reducing clock to x12. This results in improperly adjusted clock -> wrongly divided FSB -> mem clock. I need to check this is more detail as your situation seems to be a bit special, which causes the confusion in HWiNFO.

Mumak wrote on 2020-02-15, 14:11:

EIST doesn't affect FSB, the issue here is different. HWiNFO properly calculates the actual clock (4500Mhz), but then thinks this rate is based on max clock (x14) and needs to be adjusted due to EIST reducing clock to x12. This results in improperly adjusted clock -> wrongly divided FSB -> mem clock. I need to check this is more detail as your situation seems to be a bit special, which causes the confusion in HWiNFO.

Thank you for looking into it! cheers!

dottoss wrote on 2020-02-15, 14:39:

Mumak wrote on 2020-02-15, 14:11:

EIST doesn't affect FSB, the issue here is different. HWiNFO properly calculates the actual clock (4500Mhz), but then thinks this rate is based on max clock (x14) and needs to be adjusted due to EIST reducing clock to x12. This results in improperly adjusted clock -> wrongly divided FSB -> mem clock. I need to check this is more detail as your situation seems to be a bit special, which causes the confusion in HWiNFO.

Thank you for looking into it! cheers!

OK, I think I found the bug. It shall be fixed in the next build.

Slashzero wrote on 2020-02-12, 18:50:

Mumak wrote on 2020-02-12, 15:35:

Can you try to install some newer/latest version of IE if that will fix this message?
Apparently a lot of required DLLs get installed with it and that would make the life much easier 😀
I checked a Win95B Build 1212 installation and that seems to contain the required DLLs. What's your version?

I am using WIN95B (OSR). Yesterday I installed IE4. I will try HWiNFO again tonight if I have a chance. No luck. Even though I installed IE4, it still won't run.

I have good a bad news...
Good one is - I have built a version of HWiNFO32 using VS6, which should also fix the error messages about WS2_32.dll missing and hence it might run under Win95 😀
The bad one is that due to some changes I had to make (use uncompressed logo/image files embedded), this produces a significantly bigger EXE file. Uncompressed it's 21MB, so probably not suitable to be run on W95 machines which don't have so much memory 🙁
Not sure what to do now, I tried to avoid build another special version for Win95 machines, I prefer to have a single HWiNFO32 version...

Mumak wrote on 2020-02-15, 18:24:

dottoss wrote on 2020-02-15, 14:39:

Mumak wrote on 2020-02-15, 14:11:

EIST doesn't affect FSB, the issue here is different. HWiNFO properly calculates the actual clock (4500Mhz), but then thinks this rate is based on max clock (x14) and needs to be adjusted due to EIST reducing clock to x12. This results in improperly adjusted clock -> wrongly divided FSB -> mem clock. I need to check this is more detail as your situation seems to be a bit special, which causes the confusion in HWiNFO.

Thank you for looking into it! cheers!

OK, I think I found the bug. It shall be fixed in the next build.

Super cool! Thank you very much! 😀

Mumak wrote on 2020-02-12, 15:09:

CoffeeOne wrote on 2020-02-12, 14:54:

Mumak wrote on 2020-02-12, 14:44:

Thanks! Unfortunately it looks like the values I expected are not present in that node. Perhaps it's somewhere else under HKEY_LOCAL_MACHINE or not supported at all on Win9x.

Full reg

Thanks! Unfortunately no luck either. It looks like the W9x driver doesn't provide some of the interesting data in registry.

Hello,
I switched to the other eisa computer, ....
I used still hwi32_623_4075 (like on the other 486), I hope this is OK.
I started the tool in debug mode, attachments will follow.

CPU: wrong L1 cache display
The ISA PnP card is not shown, no audio devices, no harddisks, no network cards.

CoffeeOne wrote on 2020-02-16, 16:26:

Hello, I switched to the other eisa computer, .... I used still hwi32_623_4075 (like on the other 486), I hope this is OK. I sta […]
Show full quote

Mumak wrote on 2020-02-12, 15:09:

CoffeeOne wrote on 2020-02-12, 14:54:

Full reg

Thanks! Unfortunately no luck either. It looks like the W9x driver doesn't provide some of the interesting data in registry.

Hello,
I switched to the other eisa computer, ....
I used still hwi32_623_4075 (like on the other 486), I hope this is OK.
I started the tool in debug mode, attachments will follow.

CPU: wrong L1 cache display
The ISA PnP card is not shown, no audio devices, no harddisks, no network cards.

Thanks! I see there's a couple of things I need to fix.
Could you please attach registry dump of the entire HKEY_LOCAL_MACHINE or at least HKEY_LOCAL_MACHINE\Enum\ISAPNP node?

Mumak wrote on 2020-02-16, 16:41:

CoffeeOne wrote on 2020-02-16, 16:26:

Hello, I switched to the other eisa computer, .... I used still hwi32_623_4075 (like on the other 486), I hope this is OK. I sta […]
Show full quote

Mumak wrote on 2020-02-12, 15:09:

Thanks! Unfortunately no luck either. It looks like the W9x driver doesn't provide some of the interesting data in registry.

Hello,
I switched to the other eisa computer, ....
I used still hwi32_623_4075 (like on the other 486), I hope this is OK.
I started the tool in debug mode, attachments will follow.

CPU: wrong L1 cache display
The ISA PnP card is not shown, no audio devices, no harddisks, no network cards.

Thanks! I see there's a couple of things I need to fix.
Could you please attach registry dump of the entire HKEY_LOCAL_MACHINE or at least HKEY_LOCAL_MACHINE\Enum\ISAPNP node?

adding attachment ....

CoffeeOne wrote on 2020-02-16, 16:58:

Mumak wrote on 2020-02-16, 16:41:

CoffeeOne wrote on 2020-02-16, 16:26:

Hello, I switched to the other eisa computer, .... I used still hwi32_623_4075 (like on the other 486), I hope this is OK. I sta […]
Show full quote

Hello,
I switched to the other eisa computer, ....
I used still hwi32_623_4075 (like on the other 486), I hope this is OK.
I started the tool in debug mode, attachments will follow.

CPU: wrong L1 cache display
The ISA PnP card is not shown, no audio devices, no harddisks, no network cards.

Thanks! I see there's a couple of things I need to fix.
Could you please attach registry dump of the entire HKEY_LOCAL_MACHINE or at least HKEY_LOCAL_MACHINE\Enum\ISAPNP node?

adding attachment ....

Thanks! Please try this build: www.hwinfo.com/beta/hwi32_623_4084.zip
That should fix a few things like Compaq graphics name and report the PnP devices incl. SB16 PnP.

BTW: This is a special version that has been completely recompiled in VS6. So after a long time and a lot of effort, this is the first build that might be capable to run on Windows 95 too. Unfortunately I haven't been able to test this on such system yet, so I'd be happy if there are any users willing to provide feedback.
EDIT: Updated to build 4084 to fix VS6-compiler related issue.

Mumak wrote on 2020-02-16, 17:15:

Thanks! Please try this build: www.hwinfo.com/beta/hwi32_623_4083.zip That should fix a few things like Compaq graphics name and […]
Show full quote

CoffeeOne wrote on 2020-02-16, 16:58:

Mumak wrote on 2020-02-16, 16:41:

Thanks! I see there's a couple of things I need to fix.
Could you please attach registry dump of the entire HKEY_LOCAL_MACHINE or at least HKEY_LOCAL_MACHINE\Enum\ISAPNP node?

adding attachment ....

Thanks! Please try this build: www.hwinfo.com/beta/hwi32_623_4083.zip
That should fix a few things like Compaq graphics name and report the PnP devices incl. SB16 PnP.

BTW: This is a special version that has been completely recompiled in VS6. So after a long time and a lot of effort, this is the first build that might be capable to run on Windows 95 too. Unfortunately I haven't been able to test this on such system yet, so I'd be happy if there are any users willing to provide feedback.

Thx for the update, L1 cache and the graphics card look better.

EDIT: Display of the 256kB mainboard cache would be nice, too 😁