Oh dear. Might mention it is for CPU-Z and used for system identification. With any luck, they will have heard of CPU-Z, as it is fairly defacto software these days. The key is the find a person at AMD who cares and has access to archives. Most employees probably won't want to touch the subject though. Best of luck! Sometimes if you are irritating and persistent enough, they give in (squeaky wheel gets the grease).

feipoa wrote:

Oh dear. Might mention it is for CPU-Z and used for system identification. With any luck, they will have heard of CPU-Z, as it is fairly defacto software these days. The key is the find a person at AMD who cares and has access to archives. Most employees probably won't want to touch the subject though. Best of luck! Sometimes if you are irritating and persistent enough, they give in (squeaky wheel gets the grease).

I doubt that someone at AMD still have this data book, but by chance it may not be needed.
mtx500 contacted Whartman Associates, and he was told that they could not provide the data book because of their NDA with NexGen(and now AMD). But they would be OK with an authorization from AMD, that *may* be easier to obtain.

I've added an attempt to detect the IBM 386SLC/486SLC/486BL.
Any test on one of those CPU is welcome. Thanks in advance !

http://download.cpuid.com/betas/cpuz_w95.zip

(asides of this, I'm still looking for testers with a Ti 486SXLC or a Ti 486SXL - including the 2x clock variants - and Cyrix 486SLC. Thanks !).

CuPid wrote:

I need a vacation

lol! You and me both! I'll be lucky to get one in the next 20 years, assuming I'm still breathing in and out.

CuPid wrote:

I'm looking for testers with : - UMC/OPTI 82C391/82C392, 82C481/82C482 and 82C596/82C597. The software may report erroneous info […]
Show full quote

I'm looking for testers with :
- UMC/OPTI 82C391/82C392, 82C481/82C482 and 82C596/82C597. The software may report erroneous information about chipset name & L2, but I'm interested in getting a TXT report.
- Ti SXL/SXL2 (again, I could not get mine working)

http://download.cpuid.com/betas/cpuz_w95.zip

Thanks !

I have an SXL2-66 system setup. It runs SCSI2SD with Win3.11 natively, but I can swap the SD card to run NT4. Will cpuz_w95 work on NT4 and is this sufficient for your needs, or do you need W95?

feipoa wrote:

CuPid wrote:

I have an SXL2-66 system setup. Will cpuz_w95 work on NT4 and is this sufficient for your needs, or do you need W95?

CPUZ_W95 runs fine on NT4, as I recall Franck even fixed the locking for NT4 Terminal Server which has kernel namespace support.

But benchmarks results might be skewed, unless the OS used is databased.

Now that I think about it further, will an SXL2-66 even run NT4? Or does it need a "real" 486?

feipoa wrote:

But benchmarks results might be skewed, unless the OS used is databased.

Now that I think about it further, will an SXL2-66 even run NT4? Or does it need a "real" 486?

I don't think so, Franck never mentioned this to me.

How much effort would it be to "suck it and see"? I am also wondering how much memory it has, NT4 is OK with 32MB as I booted with /MAXMEM=32 two days ago, but have not tried it with less.

Not too much effort, but it has to go in queue. I am still have your SIV + Nexgen test to do. Never enough sleep.

For my SXL2-66 / ULSI DX2-66 system, I just need to swap the SD card from the SCSI2SD device. The SD card is a clone of my IBM 5x86-133/2x system, which tends to work with most other systems provided the SCSI host controller drivers are already set to load at startup. I’ve only tried this between PCI 486 systems, never with an ISA386. The SD card clone is from NT4/W95, but unfortunately I believe the ISA SCSI card in the SXL2 system (Adaptec 1540CP) can only translate up to 8 GB, so the W95 partition doesn’t get read and won’t boot to W95. I think I can boot to Windows DOS though. The NT4 partition is around 7.7 GB. So the test for NT4 is fairly quick, unless there’s s some flag I need to have set first for an SXL CPU?

The SXL2-66 system has 32 MB of RAM. Where do I set /MAXMEM = 32?

feipoa wrote:

The SXL2-66 system has 32 MB of RAM. Where do I set /MAXMEM = 32?

You don't need to set /MAXMEM as your system only has 32MB. My system has 192MB so I need to limit it for testing reasons. For your information it's set in boot.ini and on that system I have as below.

1[Boot Loader]
2Timeout=15
3Default=multi(0)disk(0)rdisk(0)partition(3)\WNT
4[Operating Systems]
5multi(0)disk(0)rdisk(0)partition(3)\WNT="WNT V4.00 SP6 MP E:"                       /NoSerialMice
6multi(0)disk(0)rdisk(0)partition(3)\WNT="WNT V4.00 SP6 MP E: 64MB"  /maxmem=64      /NoSerialMice
7multi(0)disk(0)rdisk(0)partition(3)\WNT="WNT V4.00 SP6 MP E: 32MB"  /maxmem=32      /NoSerialMice
8multi(0)disk(0)rdisk(0)partition(3)\WNT="WNT V4.00 SP6 MP E: [VGA]" /basevideo /sos /NoSerialMice
9C:\bootsect.dos="DOS V6.22 C:\DOS" /win95dos

Well, I'll give it a shot in VGA mode at least.

Well, it seemed like it was booting fine, then BSOD. I see the SCSI card in the list of drivers being loaded, so I don't think that's it. I'm going to try to image a different system I have, which has NT4 on a 4 GB partition and Win95c on 4 GB partition, then see if this works, at least for W95.

With the smaller drive partitions, 4 GB NT4, 4 GB W2K, my SXL2-66 made it to the login prompt, and then:

And if I boot into NT 4.0 (VGA Mode), I get to the Ctrl-Alt-Del login screen, but the mouse and keyboard don't work. I have to power down.

Will NT4 even run on a Cyrix DRx2-66 or SXL2-66? Is there a missing instruction? Not sure if a fresh installation will help the situation here.

@CuPid Here's the exact BSOD I get when trying to run CPU-z 1.76 (for Win98) under Xeon X5650 :

In short, "critical exception" in vxd. Also, this wasn't a "hard lock", after hiting a key I was able to return to desktop.

PS. I tried Win95 version, still the same error.

Thanks for your efforts feipoa, but don't break everything just for a test 😒

NT4 requires a 486, and naively I'd say that as long as processor has "486" in its name, then it should support the i486 instructions 😕

@agent_x007 :
yes, I know about that problem. On Westmere the programs reads a bunch of MSR registers, and some are not supported. The VxD does not not handle the exception, unfortunately.

Is Nehalem/Bloomfield(Lynnfield), the same ?
PS. I got setup up and running.

agent_x007 wrote:

Is Nehalem/Bloomfield(Lynnfield), the same ?
PS. I got setup up and running.

yes, but they may behave differently depending on their support of the concerned MSRs.
Does your system have a partition to run Win NT/2K/XP ? if yes, what would help is to get a report of cpuz under that OS, so I'll know what MSRs are supported (and will be able to remove the unsupported ones).

Report from XP SP3 in attachment (I used latest CPU-z to make it).

agent_x007 wrote:

Raport from XP SP3 in attachment (I used latest CPU-z to make it).

Perfect, thank you.

Oh, Thank You for trying to fix this.
Here's AIDA64 html raport from Windows 98 itself (version : standalone v5.92.4300, newest one get's the same error CPU-z has; FYI : I didn't tested ones between this and current newest).

Under CPUID you can find MSR table for the working core (BIOS has all six active, but OS only uses one theread).
I hope this will make things easier on your end 😀

agent_x007 wrote:

Oh, Thank You for trying to fix this. Here's AIDA64 html raport from Windows 98 itself (version : standalone v5.92.4300, newest […]
Show full quote

Oh, Thank You for trying to fix this.
Here's AIDA64 html raport from Windows 98 itself (version : standalone v5.92.4300, newest one get's the same error CPU-z has; FYI : I didn't tested ones between this and current newest).

Under CPUID you can find MSR table for the working core (BIOS has all six active, but OS only uses one theread).
I hope this will make things easier on your end 😀

I've removed five MSR readings, hoping that I did not forget one, that should run under win98 on the Westmere now. Please let me know anyway 😀

http://download.cpuid.com/betas/cpuz_w95.zip

Did a clean install of Win95c on my SXL2-66 system. Run cpuz.exe and ended up with a blue screen: