[gmx-users] gmx kill computer?

[Harry Mark Greenblatt]

BS”D

I had a similar issue with a workstation, and it turned out to be the power supply, which is why it seemed that someone pulled the plug (at least it seems to be the PSU, since having replaced it we don’t see failures).  The problem was very intermittent, and did not require any great load.  The system never had time to write any kind of error to the disk.  *But* our supplier said that there can be other problems that give rise to this sudden death.

If you can, see if swapping the power supply helps.

Harry



On 23 Oct 2018, at 4:11 PM, Michael Brunsteiner <mbx0009 at yahoo.com<mailto:mbx0009 at yahoo.com>> wrote:


Hi,
this might be considered off-topic, but i believe there is some evidence to the contrary ... what i see is this:
I have a couple of (fairly new) workstations (specs below) and on at least two of them, on at least three differentoccasions, i recently saw the following behaviour:computer is happily running and all of sudden stops doing so, as if somebody had pulled the plug.I need to turn it back on manually (i.e. it didn't just hang, or reboot - it actually turned itself off)the ONLY common factor I see is that in each of the three cases the computer, at the time when it died,had a gromacs job running with both CPU and GPU (with the same type of card) at full load ... (the samecomputers running cp2k or orca which are also quite heavy on the CPU, but not the GPU never showed issues)

e.g.:
in /var/log/syslog there's just a normal log for a cron-job, at 12:03,then at 12:10 an edr file from a running gmx job is written to the last timethen a few minutes later the computer dies without further notice.
the next /var/log/syslog entry is from 12:31 when i turned the computer on again and it starts booting.there is also NOTHING conspicuous in any of the other files in /var/log, or in any ofthe output files or stdout/stderr from the running gmx job.
the computers are NOT overclocked, cooling works, cpu temperatures are well below max.
as stated above something like this happened three times, each time with gmx at full blastand with the same type of nvidia card ... can it be that my hardware has some issue that isso well hidden that only gromacs running at full blast can trigger it? - has anybody observedsuch behaviour before, or suggestions on how to better diagnose the issue ... ??

I'd be grateful for any help!
thanks, michael

the specs:
workstation 1Intel(R) Core(TM) i7-4930K CPU @ 3.40GHz + nvidia geforce gtx 1060
running debian stretch with nvidia driver version 387.26
workstation 2AMD Ryzen 7 2700X Eight-Core Processor + nvidia geforce gtx 1060
running debian stretch with  nvidia driver version 390.87




=============================== Why be happy when you could be normal?
--
Gromacs Users mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org<mailto:gmx-users-request at gromacs.org>.


--------------------------------------------------------------------
Harry M. Greenblatt
Associate Staff Scientist
Dept of Structural Biology           harry.greenblatt at weizmann.ac.il<mailto:harry.greenblatt at weizmann.ac.il>
Weizmann Institute of Science        Phone:  972-8-934-6340
234 Herzl St.                        Facsimile:   972-8-934-3361
Rehovot, 7610001
Israel