SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Discussion about old PC hardware.

Re: SIV support for 3dfx Voodoo 3 when it's not the Primary GPU

Postby red-ray » 2019-10-17 @ 19:55

Thank you for your excellent testing, I expected it to work as I tested it on my Voodoo 3 2000, but what I could not check is that things would be OK when the VBIOS had not set I/O Enabled in the first place as I can't get my system to post with a PCI primary GPU.

I just need to fix the Voodoo Banshee memory size reading from the chip and then plan to release 5.43 Beta-00 which will include this I/O Access fix. I was expecting I would need to release Beta-00 to support the next Windows 10 20H1 insider build but M$ seem to be running late.

xjas wrote:It's still showing "0MB 64-bit SDRAM" for the Rage, but looking pretty good otherwise.

The 0MB is down to SIV not having a VXD for Windows 9x and on Windows NT/2K/XP it should get the size correct.

Good luck with sorting out the driver issues.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby DosFreak » 2019-10-17 @ 21:42

Is it possible for the interface to report that some values could not be reported due to the OS being used or is that too much work?
User avatar
DosFreak
l33t++
 
Posts: 10520
Joined: 2002-6-30 @ 16:35
Location: Your Head

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-17 @ 23:37

With and infinite timescale and budget I have been lead to believe all things are possible!

DosFreak wrote:Is it possible for the interface to report that some values could not be reported due to the OS being used or is that too much work?

Thank you for your post, I have pondered this and I am unclear as to what you would like. There are things that SIV could report on Windows 9x/Me that it does not as back in 2001 I did not design SIV for Windows 9x/Me as I expected their days were numbered. As SIV was designed for Windows NT 4.00 and later when running of Windows 9x/Me will only report a subset of what it will report on NT 4.00 and later, further running on NT 4.00 and later without admin rights will also limit what SIV can report. I am also unclear as to which OS you would like this for, my guess is Windows 9x/Me or would you also like this for things that work/are possible on Windows 10, but not on Windows 7? Further on systems with > 32 CPUs the SIV32s may/will have issues that the SIV64s will not.

Given "DosFreak" I feel I should make it clear I have no plans for a SIV that could be run from DOS :wink:

If a specific facility is not available then in general the panel will not be in the menus, but for such as Menu->Hardware->NUMA Nodes this will be available on all >= Windows XP systems, but on XP if SP3 is not installed it will report the missing kernal32 functions. Windows 9x/Me are limited to 255 controls per panel, some SIV panels need rather more than this and will report that the panel has been truncated. I have adjusted several panels such that they are different and use < 256 controls on Windows 9x/Me.

Your question is an open ended question and I suspect it's impossible to answer it fully. I feel for me to sensibly comment you need to give an example of what SIV currently reports and what you have in mind for a given OS.

Are in a roundabout way asking for the SIV support for Windows 9x/Me to be as good as the support for Windows NT 4.0 and later? I have no plans to implement this as to do this I would need to use a VXD and as Windows 9x/Me don't support such as the SEHs (Structured Exception Handlers) that the SIV design assumes will be available.

Which of your systems do you have issues running SIV on and/or what it reports? Can you post screen shots and ideally SIV save files?

In summary your question is very easy to ask and pretty near impossible to answer.
Last edited by red-ray on 2019-10-18 @ 10:31, edited 2 times in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby DosFreak » 2019-10-17 @ 23:47

No conspiracies or motives (Although SIV for DOS would be awesome ;) ) instead of having to respond when people report an invalid value then mabye make it easily recognized by the user that the value couldn't be queried due to the OS being used or perhaps permissions (Question mark with a tooltip?) and SIV doesn't support that function.

Posts similar to this that you've had to respond several times for:
The 0MB is down to SIV not having a VXD for Windows 9x and on Windows NT/2K/XP it should get the size correct.


In my job I just hate having to do the same thing over and over again and try to get rid of the problem so I don't have to so I was wondering if something like this would be feasible so you wouldn't have to answer the same question over and over.
User avatar
DosFreak
l33t++
 
Posts: 10520
Joined: 2002-6-30 @ 16:35
Location: Your Head

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-18 @ 00:37

DosFreak wrote:
The 0MB is down to SIV not having a VXD for Windows 9x and on Windows NT/2K/XP it should get the size correct.


In my job I just hate having to do the same thing over and over again and try to get rid of the problem so I don't have to so I was wondering if something like this would be feasible so you wouldn't have to answer the same question over and over.

Me too, In my initial post I feel I strongly intimated SIV on Windows 9x/Me is suboptimal and have further refined this.

red-ray wrote:
    Systems and GPUs I would like to test on
  1. SIV is designed for Windows NT so if you have both Windows NT/XP and 9X use the NT/XP system as more will be reported. Chances are that if something is zero on 9x/Me the correct value will be reported on NT 4.00 and later.
  2. As Windows 9X does not support structured exception handlers (__try { … } _except() { … }) in VXDs SIV can't sensibly use a VXD to read MSRs so struggles with Pentium III and later CPUs on Windows 9X.

In terms of tooltips I have found these are rarely if ever noticed.

To do what you seem to be asking for would be more work than redesigning SIV to work on Windows 9x/Me and makes me feel I should simply drop 9x/Me support, should I?

I look forward to hearing how SIV does on any of your systems.
Last edited by red-ray on 2019-10-18 @ 00:46, edited 2 times in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby DosFreak » 2019-10-18 @ 00:40

That's what I was wondering in my initial post if it would be too much work. Sounds like it. Thanks for working on this program and for adding support for older hardware.
User avatar
DosFreak
l33t++
 
Posts: 10520
Joined: 2002-6-30 @ 16:35
Location: Your Head

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-18 @ 01:25

DosFreak wrote:That's what I was wondering in my initial post if it would be too much work. Sounds like it. Thanks for working on this program and for adding support for older hardware.

I guess it comes down to benefit and effort. Most SIV users use it on Windows 10 x64 and that SIV works well on NT 4.00 at all is down to me having a collection of almost all the systems I have ever owned. I have been keen on MP systems since the 70s when I pined after an 11/74 MP and most of my collection are MP systems which is another reason for the lack of full 9x/Me support.

I am still hoping for a post of SIV running on an ALR 6x6.

Image
Last edited by red-ray on 2019-10-18 @ 10:16, edited 3 times in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby xjas » 2019-10-18 @ 02:02

red-ray wrote:I am still hoping for a post of SIV running on an ALR 6x6.

You might want to PM Luckybob for that, although I don't know if he ever got his working. :) The coolest SMP system I have is an ALR 7200 dual-P3/550, but it's just a board and is stuffed in a box in my storage closet at the moment.
User avatar
xjas
l33t
 
Posts: 2213
Joined: 2015-9-07 @ 02:29

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-18 @ 09:18

luckybob wrote:I will happily help out when I get "around to it".

xjas wrote:You might want to PM Luckybob for that, although I don't know if he ever got his working. :) The coolest SMP system I have is an ALR 7200 dual-P3/550, but it's just a board and is stuffed in a box in my storage closet at the moment.

I did back in August and as yet I guess he has not got "around to it", do you know or any other members with systems that have more than four sockets? I don't have the SIV save files from an ALR 7200, do you have any plans to get it running?
Last edited by red-ray on 2019-10-18 @ 10:07, edited 3 times in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-18 @ 10:04

Banshee_2d_spec.pdf wrote:When using SGRAMs, mctl_type_sdram should be set to 0. When using SDRAMs, only 16Mbit (16x512K parts) are supported, which result in a 16MB frame buffer. The sgram_type and sgram_chipsets bits in dramInit0 are ignored when mctl_type_sdram=1.

Sedrosken wrote:It does have 16MB. HWiNFO used to work but now seems to crash. Hmm. I'll have to troubleshoot that.

I checked the datasheet and have adjusted SIV 5.43 Beta-00 in light of this. Please will you check it out and post either new save files or Menu->Machine->GPU Info?

I expect HWiNFO v6.13-3955 will also report 8MB, did you manage to get It working?

Did you install the DBGHELP.dll I attached to https://www.vogons.org/viewtopic.php?f=46&t=68756&start=180#p794485?
Last edited by red-ray on 2019-10-18 @ 13:38, edited 1 time in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby Sedrosken » 2019-10-18 @ 11:34

red-ray wrote:Edit: I just spotted such as Menu->Help->SIV Lookup did not report the image type and versions. This is down to DBGHELP.dll not being there and if you put the attached into C:\WINNT\system32 then SIV should do better. I suspect you must have put PSAPI.dll V5.00.2134.1 there, did you?


Not that I know of. It's a relatively fresh install of NT4 that unfortunately it's looking like I'm going to have to re-do again for reasons outside the scope of SIV (I managed to royally screw up my drive order and can't seem to set it right again by normal means). It's going to have to wait until at least Monday, possibly Tuesday for me to be able to take a look at it. I've got a busy next couple days ahead.
The following are Pastebin links to see the specs of each named machine.
Al-Jalima, Arwic, Glenden Wood and Marae Lassel.
User avatar
Sedrosken
Member
 
Posts: 171
Joined: 2013-5-13 @ 13:53
Location: The Sticks

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-18 @ 13:42

Sedrosken wrote:Not that I know of.

OK and good luck with the drive letters, but as the system is on C: and that looks OK I am surprised that Disk Manager can't sort out things.

I have just released Beta-00 which should fix the Banshee memory size and is attached to https://www.vogons.org/viewtopic.php?f=46&t=68756
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class CPUs and 3Dfx + Matrox GPUs

Postby red-ray » 2019-10-18 @ 16:06

Sedrosken wrote:T2130CS (486DX4-75, 24MB RAM, 960K Chips&Technologies VGA) with 95A, and both CPU-Z for Win95 and SIV32L cause the machine to completely lock up. I wonder what it's angry about?

I just spotted this post from 2019-8-22 @ 21:48 and I wonder, does the latest SIV32L still trigger the lock up? Since you posted this I adjusted SIV to work on i386 and suspect that fix may also apply to your system.

If from a command prompt you do SIV32L -DBGINI > SIV_DBGOUT.log | MORE then the log should help in figuring out what may be causing any lockup.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for S3 Trio GPUs - Help needed with testing

Postby red-ray » 2019-10-19 @ 10:27

I found an S3 Trio64V2/DX in a box and after reseating the memory chips managed to get it working so added support to SIV.

I also added support for the Trio32 + Trio64 + Trio64V+ which I can't test, so if you have any of these cards it would be great if you posted a screen shot similar to as below.

I have only tested on NT4, but things should also work on W9x/WMe/W2K/WXP. The panel is Menu->Machine->GPU Info.

Image

I plan to add support for other S3 GPUs and if you would like to help post a screen shot of what SIV 5.42 Trio-01 currently reports.
Attachments
RDO.png
SIV reporting an S3 Trio64V2/DX
Last edited by red-ray on 2019-10-20 @ 23:36, edited 1 time in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby Vynix » 2019-10-19 @ 16:40

I do have a S3 Trio64V2-DX on my P133 machine, but unfortunately it is non-functional (one of the VRMs overheats and starts behaving erratically) at the moment :(

In the meantime, as you may know, I have the Omnibook with a SavageIX already, so that's already something I could test if SIV recognizes better the SavageIX (iirc it didn't report much, I'll need to have a look at the log files I previously posted :roll:).

Perhaps if I can get the P133 to boot up, I'll try to get a report on it, but I don't guarantee if it's going to work.
Proud owner of a Shuttle HOT-555A 430VX motherboard and two wonderful retro laptops, namely a Compaq Armada 1700 [nonfunctional] and a HP Omnibook XE3-GC [fully working :p]
User avatar
Vynix
Member
 
Posts: 255
Joined: 2018-12-11 @ 11:06
Location: France

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-19 @ 16:56

Vynix wrote:I do have a S3 Trio64V2-DX on my P133 machine, but unfortunately it is non-functional (one of the VRMs overheats and starts behaving erratically) at the moment :(

In the meantime, as you may know, I have the Omnibook with a SavageIX already, so that's already something I could test if SIV recognizes better the SavageIX (iirc it didn't report much).

Thank you for your post and it would be good to confirm the SIV S3 Trio64V2/DX support works on Windows 9X, but don't risk the system just to test SIV. I wonder why the VRMs overheat :( .

I have not really changed the S3 Savage/IX 86C294 support so not much will change. What other things do you feel SIV should/could report?
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby Vynix » 2019-10-19 @ 17:33

So, the mobo (Shuttle HOT-555A), has 3 VRMs, and only one of them is heating up more than it should (to the point where it becomes scorching hot) however, it only does that when I set the motherboard to use a split-voltage plane CPU, such as a Pentium MMX, if I set the motherboard to use, says, a regular Pentium (such as the P133 that came with the mobo), the offending VRM only becomes slightly warm to the touch, but otherwise stays relatively cool.

Also I think that's because it doesn't have any heatsink on it (only a thick ground trace that serves as one, well, is supposed to help with the heat), so I thought that I should put one, but that's going to require me desoldering it, bending its legs slightly to accommodate the heatsink and soldering it back (the nuclear option). Whereas if I can get one of these stick-on heatsinks and stick it on the VRM (while avoiding to short the legs)... I'm debating whether doing this or at least putting a small fan... I know someone who had a Cyrix M2 system that suffered from the same issue (the VRM would start dropping in and out), and putting a small fan to blow air on it seemingly solved the problem, and since I have a bunch of old case fans in my box of random parts and other junk, I could try and wire one to a molex harness (molex to fan adapter actually). I've done that before when I needed more fans but had no free fan headers.

As for the Omnibook, I don't know what can be done to it, it's up to you if you need to test something else, but this made me realize something: there are some PCMCIA to PCI expansion chassis, and if I could get one of them (they are by no means cheap or common but I found one on ebay for 250€ :dead: but I had the intention to get one to do some experiments with various PCI cards), stick the S3 Trio in it and see how it would go. So that's a possibility in case that the Shuttle HOT-555A soils its pantaloons at the last minute.

So all hope is not yet lost, the P133 won't be risked, I have a few tactics to counter its attempts at thwarting our plans.
Proud owner of a Shuttle HOT-555A 430VX motherboard and two wonderful retro laptops, namely a Compaq Armada 1700 [nonfunctional] and a HP Omnibook XE3-GC [fully working :p]
User avatar
Vynix
Member
 
Posts: 255
Joined: 2018-12-11 @ 11:06
Location: France

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby xjas » 2019-10-19 @ 19:11

Update: I finally got Hiren's Mini-XP to boot on the XPS T500 (for some reason my Macbook just wasn't burning the disc correctly, despite beiing able to burn every other ISO in existence just fine. >:( ) Result is attached.

SIV_XPS T500_MiniXP.7z
(132.58 KiB) Downloaded 3 times

Unfortunately I was hoping this would just be a 'just boot the CD & test' solution, but it doesn't seem to work that way. Mini-XP uses a bunch of generic drivers for hardware (such as the GPUs) so it doesn't show any info. Oh well.

red-ray wrote:I did back in August and as yet I guess he has not got "around to it", do you know or any other members with systems that have more than four sockets? I don't have the SIV save files from an ALR 7200, do you have any plans to get it running?

The ALR 7200 board is a bit buried, but I'm constantly pulling stuff out of there and shuffling things around, so next time it's accessible I'll grab it. I've been trying to sell it for ages but I'll make sure to get a result before I let it go.
User avatar
xjas
l33t
 
Posts: 2213
Joined: 2015-9-07 @ 02:29

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby red-ray » 2019-10-19 @ 19:52

xjas wrote:Unfortunately I was hoping this would just be a 'just boot the CD & test' solution, but it doesn't seem to work that way. Mini-XP uses a bunch of generic drivers for hardware (such as the GPUs) so it doesn't show any info. Oh well.

Thank you for checking out SIV on Mini-XP. Thinking about this I should have realised they the GPUs would not get reported, even with the Windows 10 pre-install environment you only get one GPU.

As it turns out it was a rather good test anyway as the system has a very small paging file which caused several of the SIV memory allocations to fail! As SIV kept working this means the error recovery paths worked as they should.

xjas wrote:The ALR 7200 board is a bit buried, but...

I know the feeling, I have several boxes containing spare bits and pieces some of which I had forgotten I had!

I have never really tried to get multiple GPUs working on 98 and also find W2K struggles. WXP and later in general are OK, but my TYAN S2466 is being a pain, when I install the 6th GPU it reports it does not have enough resources on both WXP and 2003 Enterprise Server :lol:

Image

I can't find an ATI Rage logo I like the look of
Attachments
RAP.png
Only 5 GPUs
Last edited by red-ray on 2019-10-19 @ 20:39, edited 1 time in total.
User avatar
red-ray
Member
 
Posts: 221
Joined: 2019-5-21 @ 07:08
Location: England

Re: SIV support for 386/486/586 class + Alpha CPUs and 3dfx + S3 + Matrox + old ATI + NVidia GPUs - Help needed with testing

Postby xjas » 2019-10-19 @ 20:38

red-ray wrote:As it turns out it was a rather good test anyway as the system has a very small paging file which caused several of the SIV memory allocations to fail! As SIV kept working this means the error recovery paths worked as they should.


Heh, yeah I guess stuffing a page file in a RAM drive on a 256MB system would make things pretty cramped. I'm surprised it's able squish itself in & run as well as it does.

I thought this might be interesting - here's a result from my "3-way" (sort of) SLI system running 32-bit Windows 7 and an unofficial PAE patch with 6GB RAM (+ 1.5GB VRAM.) It seems to report everything correctly. BTW I really like the realtime CPU & GPU usage and power draw monitors on the main screen. I've only been running it on retro systems so far, so I didn't realize it could do that.

SIV_Centauri.7z
(222.94 KiB) Downloaded 5 times
User avatar
xjas
l33t
 
Posts: 2213
Joined: 2015-9-07 @ 02:29

PreviousNext

Return to General Old Hardware

Who is online

Users browsing this forum: No registered users and 38 guests