Mumak wrote on 2023-07-18, 19:42:

I suppose the CPU clock is wrong too and should be 120 MHz ?

Exact

Mumak wrote on 2023-07-18, 19:42:

What SCSI controller is there?

The one on my sound card, a Media Vision Pro Audio Spectrum 16.
On windows, the SCSI part of it identify as : Trantor Media Vision SCSI Host Controller

I don't know why the CPU clock is calculated wrong but please try this version if it detects the PAS16 properly

I tried HWINFO on my 25MHz 386 a few months ago but sad to say it misbehaved badly. I haven't read the rest of this topic to see what I can do about it (sorry!)

Also, on the normal supported HWINFO32, it doesn't seem to see any IDE drives (only tested under Win98 so far).

Robbbert wrote on 2023-07-19, 09:27:

I tried HWINFO on my 25MHz 386 a few months ago but sad to say it misbehaved badly. I haven't read the rest of this topic to see what I can do about it (sorry!)

Cannot give any advice without knowing more what was wrong.. LOG+DBG would be useful.

Robbbert wrote on 2023-07-19, 09:27:

Also, on the normal supported HWINFO32, it doesn't seem to see any IDE drives (only tested under Win98 so far).

You might need to enable "Low-level IO Access" Drive Scan in HWiNFO32 there.

Mumak wrote on 2023-07-19, 07:21:

I don't know why the CPU clock is calculated wrong but please try this version if it detects the PAS16 properly
HWINF623.ZIP

* Sound card is now detected correctly!
* Keyboard still die when I go into Peripherals Information.. Note that it say "Keyboard Controller: N/A" at the last line of that section. Could it fsck something while probing it?

Rav wrote on 2023-07-19, 16:19:

* Sound card is now detected correctly!
* Keyboard still die when I go into Peripherals Information.. Note that it say "Keyboard Controller: N/A" at the last line of that section. Could it fsck something while probing it?

Thanks for the feedback, happy to see that PAS16 detection works now.
Probably some detection method in SIO or KBC is causing the problem with keyboard. Try to also disable the "Check Keyboard Controller and Super IO" option.

Mumak wrote on 2023-07-19, 09:31:

Robbbert wrote on 2023-07-19, 09:27:

Also, on the normal supported HWINFO32, it doesn't seem to see any IDE drives (only tested under Win98 so far).

You might need to enable "Low-level IO Access" Drive Scan in HWiNFO32 there.

Tried that, it worked - thanks. I also noticed that auto-update (and manual update) doesn't work on Vista and older, although I guess you already know about that. Oddly, on an old NT 4.0 machine HWINFO32 wouldn't start until I copied over the 2 files needed for update checking - even though it doesn't work anyway.

In the update error message there's a spelling mistake - you might like to fix it for the next release. It says:

Failed to retrive version information ...

retrive should be retrieve.

EDIT: just noticed what might be a bug. After starting HWINFO32 current version, the computer name is highlighted, but on the right-hand side, the Description field is blank for everything. If I click on another choice on the left and come back, then it works.

Robbbert wrote on 2023-07-23, 11:18:

Tried that, it worked - thanks. I also noticed that auto-update (and manual update) doesn't work on Vista and older, although I […]
Show full quote

Mumak wrote on 2023-07-19, 09:31:

Robbbert wrote on 2023-07-19, 09:27:

Also, on the normal supported HWINFO32, it doesn't seem to see any IDE drives (only tested under Win98 so far).

You might need to enable "Low-level IO Access" Drive Scan in HWiNFO32 there.

Tried that, it worked - thanks. I also noticed that auto-update (and manual update) doesn't work on Vista and older, although I guess you already know about that. Oddly, on an old NT 4.0 machine HWINFO32 wouldn't start until I copied over the 2 files needed for update checking - even though it doesn't work anyway.

In the update error message there's a spelling mistake - you might like to fix it for the next release. It says:

Failed to retrive version information ...

retrive should be retrieve.

EDIT: just noticed what might be a bug. After starting HWINFO32 current version, the computer name is highlighted, but on the right-hand side, the Description field is blank for everything. If I click on another choice on the left and come back, then it works.

The problem with automatic update is most likely a HTTPS certificate failure. Are you normally able to access https://www.hwinfo.com/ver.txt there?
Thanks for the notice about typo, it will be fixed in the next build.
The issue with right pane was already fixed in the latest Beta version 7.51.

Tested for ver.txt, although I believe it would be the same as trying to access the main site. On XP, it works with Firefox, but not with IE8. On Windows 98SE, not accessible in any way.

Since (from what I've seen) HWINFO uses IE technology backend to access your site, it's only going to work for Windows 7 and later. With IE no longer being supported (well, not officially), eventually that method won't work at all.

The easiest way to support all windows versions would be to use http for that, then you'd never need to worry about certificates, but of course there's the worry about so-called security possibilities. Or, alternatively, build your own updater access code. The reason to suggest http is because the OS's older than W2K have minimal or no https support. If you have it working in some way please let me know how.

I tried HWINFO 6.2.3 on my unnamed 386 computer @ 25MHz.

It's more stable than 6.1.2 was, but still crashes on the Peripherals screen, saying it can't continue, then freezing the computer. Because Ctrl+Alt+Del doesn't work, I have to press Reset.

Other things: Although it sees the Sound card in the Peripherals screen, its use of IRQ and DMA doesn't show in those screens. Also, it doesn't see my network card at all (does it check for that?).

Log and debug attached.

Robbbert wrote on 2023-07-24, 06:13:

I tried HWINFO 6.2.3 on my unnamed 386 computer @ 25MHz. […]
Show full quote

I tried HWINFO 6.2.3 on my unnamed 386 computer @ 25MHz.

It's more stable than 6.1.2 was, but still crashes on the Peripherals screen, saying it can't continue, then freezing the computer. Because Ctrl+Alt+Del doesn't work, I have to press Reset.

Other things: Although it sees the Sound card in the Peripherals screen, its use of IRQ and DMA doesn't show in those screens. Also, it doesn't see my network card at all (does it check for that?).

Log and debug attached.

You might need to disable the "Check Keyboard Controller and Super IO" option in HWiNFO to avoid freezing on peripherals.
I can see it found the "SMC EtherCard PLUS Elite16 (SMC 8013W) LAN Adapter" and also reported the port and MAC address.
What sound card do you have?

I disabled the "Check keyboard..." thing, this time no crash.

I didn't open the computer, but afaik the card is a standard old Sound Blaster 16 with manual jumpers. HWINFO says it's a Creative Sound Blaster 2.0.

Hitting enter on the description produced a new bug. First time it said DSP 2.0, I/O 220 and nothing else. Pressing Esc and trying again now produced nonsense, and each time after makes different rubbish. For example it said DSP 33.255, with 4 IRQs listed, 3 8-bit DMAs and 3 16-bit DMAs. Each time after changes the DSP to a random number.

Despite all that, if I then go back to the IRQ and DMA screens, there's no sign of the sound card still, the actual assigned ones showing as free.

Robbbert wrote on 2023-07-24, 12:39:

I disabled the "Check keyboard..." thing, this time no crash. […]
Show full quote

I disabled the "Check keyboard..." thing, this time no crash.

I didn't open the computer, but afaik the card is a standard old Sound Blaster 16 with manual jumpers. HWINFO says it's a Creative Sound Blaster 2.0.

Hitting enter on the description produced a new bug. First time it said DSP 2.0, I/O 220 and nothing else. Pressing Esc and trying again now produced nonsense, and each time after makes different rubbish. For example it said DSP 33.255, with 4 IRQs listed, 3 8-bit DMAs and 3 16-bit DMAs. Each time after changes the DSP to a random number.

Despite all that, if I then go back to the IRQ and DMA screens, there's no sign of the sound card still, the actual assigned ones showing as free.

Hmm, that looks like the communication with the SB is not stable enough.. Would be interesting to get a new LOG + DBG of that page, now that it doesn't hang there due to SIO check.
BTW, I will also add a workaround to skip the problematic part of SIO check on some boards to avoid such issues.

Here's the latest info. Today when I was in the debug mode, when I went to exit the Peripherals screen I had a continual flashing of Can't Continue with the box opening and closing nonstop by itself. I hit the Reset button to regain control.

Current version of HWiNFO for DOS:
1. Doesn't detect secondary display. But it is possible (as I test with Geforce 2/3/4 cards with DVI+VGA ports). When you query DDC information from display you should do it like this, manipulating child_id:

1ax = 0x4F15 (VESA DDC API)
2bx = 0x00,0x01, ... etc (VESA DDC function select)
3cx = 0, 1, 2, ... etc (child_id - display device number)
4int 0x10

2. Doesn't detect secondary video adapter. F.e. a hercules-compatible ISA board with primary VGA. It is possible to do by pokeing 0x3Bx (for mono) ports and vice versa 0x3Dx (for color). And reading BIOS DATA AREA at segment 0x40 could show you which device is primary - color or mono.
3. Also when you read/write video memory even in text mode you could compare results in 8-16-32bit reads/writes and detect video card bus width. So you could guess either ISA8, ISA16, VLB or PCI. Some ISA16 cards could be configured by jumpers for either ISA8 or ISA16 transfer mode. VIDSPEED video test by John Bridges uses this.

b_w wrote on 2023-08-05, 11:16:

Current version of HWiNFO for DOS: 1. Doesn't detect secondary display. But it is possible (as I test with Geforce 2/3/4 cards w […]
Show full quote

Current version of HWiNFO for DOS:
1. Doesn't detect secondary display. But it is possible (as I test with Geforce 2/3/4 cards with DVI+VGA ports). When you query DDC information from display you should do it like this, manipulating child_id:

Copy code to clipboard

1ax = 0x4F15 (VESA DDC API)
2bx = 0x00,0x01, ... etc (VESA DDC function select)
3cx = 0, 1, 2, ... etc (child_id - display device number)
4int 0x10

2. Doesn't detect secondary video adapter. F.e. a hercules-compatible ISA board with primary VGA. It is possible to do by pokeing 0x3Bx (for mono) ports and vice versa 0x3Dx (for color). And reading BIOS DATA AREA at segment 0x40 could show you which device is primary - color or mono.
3. Also when you read/write video memory even in text mode you could compare results in 8-16-32bit reads/writes and detect video card bus width. So you could guess either ISA8, ISA16, VLB or PCI. Some ISA16 cards could be configured by jumpers for either ISA8 or ISA16 transfer mode. VIDSPEED video test by John Bridges uses this.

Yes, HWiNFO for DOS doesn't detect such multiple PCI adapters. That is a rather "modern" feature, not well suitable to be detected by a DOS tool.
It can however detect some special accelerator cards (e.g. ACRTC, TIGA, SPEA, IGC20, ..).
HWiNFO doesn't check MDA 3Bx ports for such adapters but if there are multiple adapters it might be a problem. I'd need to see the LOG+DBG file.

"doesn't detect such multiple PCI adapters" ? You maybe don't understand me. In pp. 1. I mean when two or more monitors connected to ONE videocard. F.e. some Matrox/ATI/nVidia cards have two DVI or two VGA output ports or either VGA+DVI. This detection code is rather simple. I could provide any logs needed to resolve these detection issues. 😀

b_w wrote on 2023-08-05, 16:20:

"doesn't detect such multiple PCI adapters" ? You maybe don't understand me. In pp. 1. I mean when two or more monitors connected to ONE videocard. F.e. some Matrox/ATI/nVidia cards have two DVI or two VGA output ports or either VGA+DVI. This detection code is rather simple.

Ah yes, sorry.

Hiya! Thanks for the tool, it's cool playing with it. I have enclosed reports on the two computers I currently have set up that are able to run plain DOS, should you still be looking for them. I also included a HWiNFO32 report from one of the two that also has Windows 98 SE installed.

Of the machines, one is a P3-733 I've built from whatever parts I've managed to gather (for example, the motherboard was yanked from a crappy Fujitsu-Siemens computer), the other one is an IBM Aptiva from 1995... although not quite in its original setup, as I had to replace the failing original optical drive with a newer one, and I've also added more memory to it.

The P3-733 had some (very mild) weirdness going on in HWiNFO:

^ It's supposed to be 733 MHz. Though that's not far off; I suppose you still calculate this value in the code?

^ This weirdness sometimes happens; it shows two ATAPI devices on the secondary IDE, although there is supposed to be one hard drive and one optical drive, like in the next image:

^ This is what I actually have in there.

^ ...and from what I can tell, it does detect & display the Maxtor hard drive (that is, the master on the secondary IDE channel) quite correctly when it feels like noticing its existence.

---

I didn't snap any photos of the Aptiva, but it sometimes also showed two ATAPI devices on an IDE channel that has a master hard disk and a slave optical drive. Oh, and it also detects the Vibra 16 as a plain Sound Blaster 16; I dunno if there even is a way to tell the two apart, and as it's the only ISA sound card I own, I cannot really do any further testing.

Thanks for the feedback.
Yes, the CPU clock is measured using TSC here, such small inaccuracy is expected under DOS given such (relatively) high clocks.
For the invalid ATAPI device, what drive details does it show in that case? Is there some other mismatch in the info?
The ViBRA16 can be detected only if it appears as a PnP device which doesn't seem to be the case here:
PnP Capable Devices: None