consistent error, but only one test (6) and only w/ parallel CPU -- real problem?

[shubckin]

YES! with 4.3.5b test№6 with parallel mode passed on this computer:
Desktop gigabyte-z77-d3h Intel sandy bridge i7-2700K with 16gb (4x4gb) of ram
On this week i will run all test on all computers in this parallel mode and write to this thread.
Thank you gyes!

[keith]

With 4.3.5a problem still remained

We may have fixed the issue, can you try this 4.3.5b build:

USB image:
http://www.passmark.com/ftp/memtest86-usb-4.3.5b.zip

CD image:
http://www.passmark.com/ftp/memtest86-iso-4.3.5b.zip

[shubckin]

With 4.3.5a problem still remained

[shubckin]

ok i will try

[David (PassMark)]

Can you try this pre-release build and see if the problem still occurs:

USB image:
http://www.passmark.com/ftp/memtest86-usb-4.3.5a.zip

CD image:
http://www.passmark.com/ftp/memtest86-iso-4.3.5a.zip

[shubckin]

bug with test 6 in parallel mode

Hi all!i have same problem with 4.3.4 LAST VERSION on 3 computers:
1. Server Supermicro sys-5017R-mtf with Xeon-2620 32gb mem (4x8gb)
2. Desktop gigabyte-z77-d3h Intel sandy bridge i7-2700K with 16gb (4x4gb) of ram
3. Desktop gigabyte-z77-ds3h Intel ivy bridge i5-3330 with 8gb (2x4gb)
On all computers with parallel mode on test 6 it FREEZES and no response to keyboard or anything. It also freezes with 1 stick of ram in parallel mode
Help? how to solve this problem!
Thanx!

[David (PassMark)]

Version 4.3.4g, V4.3.4 & the the V4.3 that comes with V5 are all more or less the same software.

You are saying it crashes in single CPU mode, multi-cpu mode, going into the configuration window and setting values in the configuration window. Basically it is crashing all the time for you.

Given that we have no other reports of a similar fault during the month since your initial post, it must either be some issue specific to the E5-2620, or a fault with your machine.

Would it be possible to test with 1 stick of RAM and 1 CPU?

[AlexVB]

Sorry for delay. I had no access to that machine for some time.

1) As for cpu's: we are using 2 Intel Xeon E5-2620 processors
2) As for crashes:

in the version memtest86-iso-4.3.4g.zip (27 sep):
1) start the program
2) press c (configuration)
3) press 4 (cpu selection mode)
4) press 1 (parallel (all))
5) press 0 (continue) - system resets, every time

in the latest version 4.3.4 (1 oct)
1) random reboots when i press c (configuration)
2) random crashes when i select multi cpu (like above but random)
3) rendom crashes in random test in single cpu mode (system just reboots)

in the version 4.3 that comes with version 5
UPDATE:
1) sometimes when i select multi cpu mode (like above), system reboots, but not every time.

And also we have no any problems with that machine, it is working fine for a weeks.

[David (PassMark)]

There is no point using V4.3.0 in my opinion as it has a known bug.

Would be better to either work out why V4.3.2 doesn't work on this machine, or use V5.

With V4.3.2, when exactly does it reboot? I assume you get as far as selecting parallel mode and then starting the tests. But in what test does it reboot? Does it always reboot in the same place in the same test?

Note also that random reboots can be a sign of bad hardware.

[AlexVB]

AlexVB,
What version of MemTest86 are you using?
What CPU are you using?

I used memtest 4.3.0, because multi-cpu mode does not working in our system in the version 4.3.2 (system just reboots when i select parallel mode).


As for CPU and screenshots - i will post that information in 1-2 days.

[David (PassMark)]

AlexVB,
What version of MemTest86 are you using?
What CPU are you using?

Subsequent to my previous post above, we found and fixed a bug in Test #6. This was fixed in Version 4.3.1 (8/Aug/2013).

We aren't aware of any definite issues in Test 7. We did have another report of errors in Test #7 in multi-threaded mode, but we are assuming this was a real hardware fault.

Can you post a screen shot showing the errors.

We are in the process of re-writing all the ECC handling code for V5. We hope to be able to report actual ECC errors for common CPUs. (As opposed to only reporting the errors that ECC failed to correct, as is the case at the moment).

[AlexVB]

We have Supermicro X9DR3-F with 24GB of ECC memory and memtest shows errors in tests 6 & 7 in multi-cpu mode. Buf if we set lower memory limit to 1G test completes without errors in multi-cpu mode. (Of course there are no errors with or without memory limit in single-cpu mode).

[David (PassMark)]

We ended up implementing a work-around in the V4.3 release that should avoid this problem.

See the post on the MemTest86 V4.3 release for more details.

[David (PassMark)]

We spent a few more days looking at the problem, but didn't make any progress on fixing it. There are some work around solutions, like turning off multi-threading for test #6, but we would prefer to fix the root cause.

If we don't solve it in a few more days, we'll probably implement a work around in V4.3 and instead concentrate on getting it to work properly in V5.

[Perbear]

I continued setting up my workstation and it seems stable, running 15 million cell CFD problems without indication of any RAM or other hardware problems.

When you have updated software I would like to test it.