Mumak wrote on 2020-03-19, 20:48:

I'd be happy to get report files from your system(s). That helps me to analyze the outcome and see if I can improve something.

Thanks for the app update. I have a few machines will test this weekend and attach reports of any that are not properly recogized.

Mumak wrote on 2020-03-19, 20:38:

Baoran wrote on 2020-03-19, 19:24:

Mumak wrote on 2020-03-19, 18:31:

Please try this version: HWINF602.zip
I expect it not to crash during video detection and Gravis check. Would be great if you could post the report file of the system.

Here is a partial log file. It freezes in "sound blaster and compatible information". It shows 3 lines there, first one is dsp 1.5, second is dsp copyright that is empty and third is port 220 and after that the system is frozen and I have to reset.

HWINFO.LOG

Thanks. This should fix the hang in SB info: HWINF602.zip
But I don't know how to determine the IRQ/DMA for older SB cards (prior to SB16).

Here is the full log file.

I think that if blaster variable does not exist or if you don't want to trust it the main way to check interrupt of a sound blaster card is to go through all the possible interrupts and see which one the DSP responds to.
On my CT1320C sound blaster you can't change the dma and it is always 1 and it can only be disabled/enabled by a jumper. When it comes to dma in general it seems most test programs (like pctools 9.0 for example) either trust the blaster variable or they assume that it is 1 and then play a digital sound and prompt the user to reply if they heard a sound or not.

keenmaster486 wrote on 2020-03-19, 04:34:

386+ is fine, of course, but a 16-bit version I would LOVE as well, if possible.

Well, considering that HWiNFO needs 2+ MB RAM, it would be useless for majority of 286 machines anyway.
Also, for 16-bit PCs Checkit and Checkit Pro are good enough.

Mumak wrote on 2020-03-19, 08:39:

Currently it's 386SX up to approx. Intel Core family, perhaps a few newer too. But later ones might lack a few things as I scrapped some support to save memory. I don't expect users of those system to be running DOS.

I don't use DOS on modern machines, but I don't use Windows on them, either.
And while it's easy to boot DOS from a diskette/CD/whatever, without installing it, it's not so easy to use Windows this way.
So yes, it would be nice to have some modern hardware info tool - not necessarily in the form of a DOS program, but maybe a bootable image which doesn't require any particular OS to be installed?

One thing I noticed was that 550Kb of conventional ram wasn't enough. I had 552Kb free and the program loaded fine, but I got out of memory error during the tests itself.

douglar wrote on 2020-03-20, 01:36:

Here's what is says on my gateway bat4ip3 mobo-- […]
Show full quote

Here's what is says on my gateway bat4ip3 mobo--

Copy code to clipboard

1                                 HWiNFO v6.0.1 
2  º Main Processor:           AMD Am5x86, 119.6 MHz                          º
3  º Bus Type:                 PCI (V2.00) + ISA                              º
4  º Mainboard Model:          Unknown                                        º
5  º Mainboard Chipset:        Intel 82420EX                                  º
6  º First Level Cache:        1 KBytes                                       º
7  º Second Level Cache:       256 KBytes of Asynchronous SRAM                º

Three oddities:
[*] The first level cache seems small.
[*] The reported speed varies between 110 and 120Mhz.

When I try and do a full report, it locks up here:
Photo Mar 19, 8 21 19 PM.jpg

The log that it did generate is attached.

Thanks. Could you please manually step thru the screens and tell me at which exact item in the ISA screen it locks up? It should be showing an indicator at bottom what it is currently checking...

Grzyb wrote on 2020-03-20, 05:50:

I don't use DOS on modern machines, but I don't use Windows on them, either. And while it's easy to boot DOS from a diskette/CD/ […]
Show full quote

Mumak wrote on 2020-03-19, 08:39:

Currently it's 386SX up to approx. Intel Core family, perhaps a few newer too. But later ones might lack a few things as I scrapped some support to save memory. I don't expect users of those system to be running DOS.

I don't use DOS on modern machines, but I don't use Windows on them, either.
And while it's easy to boot DOS from a diskette/CD/whatever, without installing it, it's not so easy to use Windows this way.
So yes, it would be nice to have some modern hardware info tool - not necessarily in the form of a DOS program, but maybe a bootable image which doesn't require any particular OS to be installed?

Why not use a WinPE bootable USB with HWiNFO32/64 for later systems?

There is some bug which reporting that HWinfo is running under Win95/98 when just MS-DOS 7.1 is used.

Otherwise i dunno if it supported, but is nice to see what virtualization tools trying to emulate..
I tested Vmware and Virtualbox and there are working fine, with Qemu so far i wanst lucky, but im using my custom build, i have to test vanilla one.

There is small Vmware problem, Video info first screen, is showing hourglass cursor cursor icon.. indefinitely, but when i press enter it could continue fine, but user could thing that he got freeze.

I tested Qemu vanilla.. There is fatal error right after start.

Mumak wrote on 2020-03-20, 07:30:

Why not use a WinPE bootable USB with HWiNFO32/64 for later systems?

So, HWiNFO does work under PE... problem solved, then.
I had problems running some stuff in PE, so I didn't even try with HWiNFO...

ruthan wrote on 2020-03-20, 08:44:

There is some bug which reporting that HWinfo is running under Win95/98 when just MS-DOS 7.1 is used. […]
Show full quote

There is some bug which reporting that HWinfo is running under Win95/98 when just MS-DOS 7.1 is used.

Otherwise i dunno if it supported, but is nice to see what virtualization tools trying to emulate..
I tested Vmware and Virtualbox and there are working fine, with Qemu so far i wanst lucky, but im using my custom build, i have to test vanilla one.

There is small Vmware problem, Video info first screen, is showing hourglass cursor cursor icon.. indefinitely, but when i press enter it could continue fine, but user could thing that he got freeze.

V86 managers do not allow low-level access to certain hardware, so this is definitely not recommended.
And you can't expect a DOS application to work properly under hardware-virtualized host like VMware, it's totally isolated.

Grzyb wrote on 2020-03-20, 08:50:

Mumak wrote on 2020-03-20, 07:30:

Why not use a WinPE bootable USB with HWiNFO32/64 for later systems?

So, HWiNFO does work under PE... problem solved, then.
I had problems running some stuff in PE, so I didn't even try with HWiNFO...

Yes, both HWiNFO32 and HWiNFO64 fully support WinPE. In some cases you might also need to load certain drivers (i.e. GPU) to get some information.

Mumak wrote on 2020-03-20, 08:55:

V86 managers do not allow low-level access to certain hardware, so this is definitely not recommended.
And you can't expect a DOS application to work properly under hardware-virtualized host like VMware, it's totally isolated.

I would expect that it will not working with some interpreter as Dosbox, but Virutalization tools are emulation whole machine, there should be difference between real one and virtual one, from programs point of view.. Except dos sound, programs are working fine, maybe better than on real machines, because its emulation typical machine (Qemu and Virtual are emulation i440fx chipsets), not something strange. Navratils system info is working fine with emulated hw from my experience.
Of course some implementation could be buggy, you can get some unexpected values, what is typical with Qemu, it has more bugs than others.

ruthan wrote on 2020-03-20, 09:14:

Mumak wrote on 2020-03-20, 08:55:

V86 managers do not allow low-level access to certain hardware, so this is definitely not recommended.
And you can't expect a DOS application to work properly under hardware-virtualized host like VMware, it's totally isolated.

I would expect that it will not working with some interpreter as Dosbox, but Virutalization tools are emulation whole machine, there should be difference between real one and virtual one, from programs point of view.. Except dos sound, programs are working fine, maybe better than on real machines, because its emulation typical machine (Qemu and Virtual are emulation i440fx chipsets), not something strange. Navratils system info is working fine with emulated hw from my experience.
Of course some implementation could be buggy, you can get some unexpected values, what is typical with Qemu, it has more bugs than others.

HWiNFO for DOS accesses the hardware at very lower level and none of the virtualization tools can fully emulate all such aspects, hardware registers and their exact behavior and timing.
Tools that rather rely on some services, i.e. software interrupts, VESA or BIOS calls (that are easily to emulate) might work better under VM, but those that access registers directly is a bit different story.
Nevertheless, HWiNFO for DOS has always been designed to work with real hardware.

HWiNFO for DOS accesses the hardware at very lower level and none of the virtualization tools can fully emulate all such aspects, hardware registers and their exact behavior and timing.
Tools that rather rely on some services, i.e. software interrupts, VESA or BIOS calls (that are easily to emulate) might work better under VM, but those that access registers directly is a bit different story.

Why you thing that all these low level thing arent emulated within this virtualization technologies? I dont thing that without it would emulation work. They are completely separated from real machine, its not some semi virtualization with passthrough etc, so any Windows / Linux / MacOSservices and real machine Bios calls etc.. doesnt matter. Imagine is a complete ZX spectrum etc emulators, here also all programs works as on real machine, if there is not some emulation layer bugs.

my system roughly;
FIC PA-2013 v2.1 motherboard, VIA MVP3, k6-2+ 500mhz. 512mb ram.
1 AGP (Matrox G400 MAX)
1 PCI (3COM 3C905C)
2 ISA (Gus PNP, AWE32 PNP)

I tested with hwinf602.ziop

mainboard : it misses my third level cache of 1mb on my motherboard, with my amd k6-2+ which has 1st + 2nd level cache on cpu. 3rd level on board.
(under PnP DMI Support page, I can see the 1mb of cache)

award bios bage: detects 2 PnP CSN's. but only lists my GUS PnP and not my AWE32.
but if I go into PnP ESCD support I see it listed in isa slot 7 it shows as AWE64 Wavetable (GUS is listed in isa slot 0)

bus info page: does not detect anything on my ISA bus (my awe32 + gus pnp are on it). Says ISA BUS "No devices/cards found"

Under Perhiperals page it lists sound card as Gus PnP but if I click on it it gives me the AWE32 info (The SB parts on my GUS PnP are disabled in flash bios). It also says 0 bytes dram when it has 32mb installed (28mb usable) but gives correct AWE32 IRQ/DMA/PORT.

I wonder if its confused by the GUS because GUS can have SB emulation. But it should still see it on the PnP CSN (I dont believe GUS enumerates the AWE32 CSN from Creative).

I can get screenshots if needed

I ran the tool on my 386. It works, but complains about EMM386, as expected. The screen is blinking all the time... I ran Dr Hardware 9 as well, as a comparison.

https://youtu.be/sQfRCzM3RD4

So something is off here...

the blinking is because I think, its trying to go into 'ice' colour mode. I dont know its real name. But it uses blink as hi-colour in text mode to give you extra background colours.

Oh that sounds reasonable. It’s probably messing up something in the Tseng‘s attribute controller. Forgets to disable blinking or such. Would have to look up in the VGA docs myself...

BloodyCactus wrote on 2020-03-20, 12:37:

my system roughly; FIC PA-2013 v2.1 motherboard, VIA MVP3, k6-2+ 500mhz. 512mb ram. 1 AGP (Matrox G400 MAX) 1 PCI (3COM 3C905C) […]
Show full quote

my system roughly;
FIC PA-2013 v2.1 motherboard, VIA MVP3, k6-2+ 500mhz. 512mb ram.
1 AGP (Matrox G400 MAX)
1 PCI (3COM 3C905C)
2 ISA (Gus PNP, AWE32 PNP)

I tested with hwinf602.ziop

mainboard : it misses my third level cache of 1mb on my motherboard, with my amd k6-2+ which has 1st + 2nd level cache on cpu. 3rd level on board.
(under PnP DMI Support page, I can see the 1mb of cache)

award bios bage: detects 2 PnP CSN's. but only lists my GUS PnP and not my AWE32.
but if I go into PnP ESCD support I see it listed in isa slot 7 it shows as AWE64 Wavetable (GUS is listed in isa slot 0)

bus info page: does not detect anything on my ISA bus (my awe32 + gus pnp are on it). Says ISA BUS "No devices/cards found"

Under Perhiperals page it lists sound card as Gus PnP but if I click on it it gives me the AWE32 info (The SB parts on my GUS PnP are disabled in flash bios). It also says 0 bytes dram when it has 32mb installed (28mb usable) but gives correct AWE32 IRQ/DMA/PORT.

I wonder if its confused by the GUS because GUS can have SB emulation. But it should still see it on the PnP CSN (I dont believe GUS enumerates the AWE32 CSN from Creative).

I can get screenshots if needed

Thanks for the extensive report. Yes, I believe it's confused by those 2 sound cards. Can you please attach the full report file so I can see the exact data and look how to fix this?