Re: Problems booting AMD64 Dual Core



On Wed, Nov 28, 2007 at 09:24:49PM +0000, I.E.Broadbent wrote:
I built two boxes last week...
both (almost) identical....
CPU - Both AMD Socket-AM2 64bit Dual Core (one 4300 and one 4600)
Memory - both identical - 4 x 1gig modules per motherboard
Motherboard - both identical - Gigabyte S-Series M61P-S3
Both using onboard Network and Graphics


One machine works fine.... no problems.... in fact superb fast and sweet
as a nut.

The other ... is a dog!
It wont output any display at all when all 4 memory slots are populated,
and when using only 1,2, or 3 it 'stutters' along working for about 40
seconds then locking up for 10 accepting no input and producing no
output.

I have exchanged the motherboard for another, making no difference
Have upgraded the BIOS to version'f' - no improvement
Have changed the power-supply (and enlarged) to ensure it's not a
voltage issue.

I even used different memory modules to see if it made a difference, ..
but it didn't.

The only thing I haven't changed is the CPU... but the supplier is
reluctant to swap it out .. claiming that if it is working at all (even
crippled) then it wont be the CPU at fault. My claim is that it is
the ONLY thing left, and if the dual-core addressing has an error then
it might explain the 'stuttering' and the failure to boot when connected
to 4 modules.

My first guess would be memory but you've changed that; no guarantee
that the new stuff is OK. Try pulling the memory from the board that
works great. Try just using one stick, then two, then four.

Compare the dmesg and /var/log/syslog from each box. Try a serial
console and the first thing you do on boot is cat /var/log/syslog. This
way you can capture it before the box crashes.

Swap drives between the systems.

Last (since you have to disturb the heat-sink grease), swap the CPUs
from one to the other MB.

Doug.


--
To UNSUBSCRIBE, email to debian-user-REQUEST@xxxxxxxxxxxxxxxx
with a subject of "unsubscribe". Trouble? Contact listmaster@xxxxxxxxxxxxxxxx



Relevant Pages

  • Re: Keys get stuck
    ... Swap can definitely keep X off the cpu for extended periods, ... A heavily swapping machine ... Seems ou use too much memory then. ...
    (Linux-Kernel)
  • Re: [patch 00/20] VM pageout scalability improvements
    ... only does it use up CPU time, but it also provokes lock contention ... and can leave large systems under memory presure in a catatonic state. ... all the CPUs in the pageout code and no actual swap IO. ... The noreclaim patches come verbatim from Lee Schermerhorn and ...
    (Linux-Kernel)
  • Re: [patch 00/20] VM pageout scalability improvements
    ... only does it use up CPU time, but it also provokes lock contention ... and can leave large systems under memory presure in a catatonic state. ... all the CPUs in the pageout code and no actual swap IO. ... The noreclaim patches come verbatim from Lee Schermerhorn and ...
    (Linux-Kernel)
  • Re: OT: memory requirements
    ... and xorg is using 70% of my memory. ... How beefy is the CPU? ... If it uses more than a few MB of swap, ... To UNSUBSCRIBE, email to debian-user-REQUEST@xxxxxxxxxxxxxxxx ...
    (Debian-User)
  • Fwd: Total vs per-cpu memory
    ... correct as there are 4GB per cpu. ... from the code it is clear that memory is ... insufficient to rotate the object in a non-parallelized part of the program. ... To UNSUBSCRIBE, email to debian-user-REQUEST@xxxxxxxxxxxxxxxx ...
    (Debian-User)