Have you tried Memtest86+ (with the plus sign)?

There's a DOS version of Memtest86+, too. If you're able to boot into DOS, you could try running the exe from there.
It's no longer available for the latest version, but older versions are still online: http://www.memtest.org/#downiso (get the "Pre-Compiled EXE file")

jwt27 wrote:

Have you tried Memtest86+ (with the plus sign)?

There's a DOS version of Memtest86+, too. If you're able to boot into DOS, you could try running the exe from there.
It's no longer available for the latest version, but older versions are still online: http://www.memtest.org/#downiso (get the "Pre-Compiled EXE file")

Thanks. Version 2.00 appears to work ("Memtest86+ v2.00").

I think memtest86+ before version 4 works.

MemTest86+ version 4.00 should work.

Had the same problems.
Use 4.00.

Hi,
MemTest86+ 4.10 did work, but not Memtest86+ 4.20
I guess they've made a change to use Pentium timers in 4.20.

https://www.memtest.org
https://en.wikipedia.org/wiki/Time_Stamp_Counter
https://en.wikipedia.org/wiki/High_Precision_Event_Timer

I have been using Memtest86+ 4.10 with 386 and newer hardware without problems.

Any benefit of 4.10 over 4.00? I've never tried 4.10.

I expect you guys know that 4.20 introduces "failsafe mode", but also there was at least one version of 4.20 that was buggy. The one that shipped with Ubuntu 12.10 LiveISO was false-positive happy. I don't know how it was limited to that release of 4.20, or if it actually extended further.

Just mentioning that in case you haven't tested newer, or just a new-ish download of 4.20. My 486 isn't set up, and I don't remember the vers I last used.

memtest86+ saved me so much trouble and builds trust from customers. Better one to catch bad memory than any software I found even the microsoft's memory tests passed it which is not good thing.

While ago, I had an customer with notebook that would boot and go BSOD, thinking either bad hard drive or software. I knew better and got memtest86+ on usb and booted with which promptly detected one bad memory module out of two and replaced it with exact match when special order came in. Fixed perfectly.

Cheers, pentiumspeed

I've been using Memtest86+ v2.11. I'm not sure how I got hold of this particular version, and it does seem to have a bug that causes it to consider a pass completed at around 120%. But that's a superficial flaw, and other than that it works with every computer I've tried it on. Might be time to go ahead and update though. 😀

feipoa wrote:

Any benefit of 4.10 over 4.00? I've never tried 4.10.

After reading through the change logs...
I don't think any past v2.11 add anything new that would help with pre-DDR2 vintage equipment's.
v4.00 and up seem to only add things for DDR2 equipment and newer so for retro I don't see any benefit in using 4.10 over 4.00.
.

Strange?
4.10 is so new that it can NOT tell if my ram is EDO or FPM?
It does not know what my chipset, cache... NOthing is.

486 LuckyStar MoBo with 4*8MB FPM?

Could be the opposite that it's so old it doesn't know. Unless another version worked for you. I doubt it matters as long as it works. nice necro btw.

Warlord wrote on 2021-08-18, 06:43:

Could be the opposite that it's so old it doesn't know. Unless another version worked for you. I doubt it matters as long as it works. nice necro btw.

Thanx, i also attached some photos in the previous post.

holoprox wrote on 2021-08-18, 06:40:

Strange? 4.10 is so new that it can NOT tell if my ram is EDO or FPM? It does not know what my chipset, cache... NOthing is. […]
Show full quote

Strange?
4.10 is so new that it can NOT tell if my ram is EDO or FPM?
It does not know what my chipset, cache... NOthing is.

486 LuckyStar MoBo with 4*8MB FPM?

Detecting the type of memory depends on the chipset. Reading chipset configuration is a feature memtest got around Pentium III time, if I remember correctly, so around 2000. No one cared about implementing getting configuration from old 486 chipsets at that time, so the feature is just not implemented. One reason for this is that for many chipsets of that time, datasheets are publicly not available, although IIRC your LS board uses a SiS 496/497 chipset which has a publicly available datasheet.

The cache size is determined by doing timing measurement, I think. memtest doesn't implement any classic timing methods, but relies on the cycle counter (time stamp counter) that is only provided from Pentium onwards, so you won't get cache size and performance data on 486 class computers.

Pretty old topic but let's pull it up. Does anybody have a working floppy image that can test the ram on my FIC PIO-2 with a dx4 which they can provide? So far I did not find any working memtest. ctramtest complains that it cannot detect my host bridge and exits, memtest86 just reboots and memtest86+ in either 4.0 or 2.x are iso's only and too large to convert them to a floppy image.

marbury wrote on 2023-03-02, 12:22:

Pretty old topic but let's pull it up. Does anybody have a working floppy image that can test the ram on my FIC PIO-2 with a dx4 which they can provide? So far I did not find any working memtest. ctramtest complains that it cannot detect my host bridge and exits, memtest86 just reboots and memtest86+ in either 4.0 or 2.x are iso's only and too large to convert them to a floppy image.

Memtest86+ runs on a 386 up until v4.10. You can find a nice collection of all floppy versions in this thread: Memtest86+ old versions

konc wrote on 2023-03-02, 12:55:

Memtest86+ runs on a 386 up until v4.10. You can find a nice collection of all floppy versions in this thread: Memtest86+ old versions

Will try them out. Thanks.