| 1 |
This is turning into a bad month :-( |
| 2 |
|
| 3 |
I'm running BOINC clients on this box, and the kernel seems unable to |
| 4 |
schedule them properly. I'm subscribed to several projects, so I should |
| 5 |
have one on each CPU all the time, running at nice 19 and therefore mopping |
| 6 |
up all available CPU cycles. That's how it used to run. But nowadays the |
| 7 |
kernel scheduler insists on allocating both of them to the same CPU, thus |
| 8 |
limiting them to 50% load. Occasionally it will start up correctly, but |
| 9 |
only if I've started the BOINC client interactively rather than from a |
| 10 |
startup script, but even if so it still reverts to its bad behaviour after |
| 11 |
a while. I haven't been able so far to spot any particular influence that |
| 12 |
might cause this reversion, and the time before it happens is apparently |
| 13 |
random. |
| 14 |
|
| 15 |
The box is a Supermicro H8DCE with 2 x Opteron 246 CPUs and 2 x 2GB RAM. |
| 16 |
This board divides the DIMM slots into two banks of four, one bank next to |
| 17 |
each CPU and associated with it. I've tried various kernels from 2.6.16-r13 |
| 18 |
to 2.6.21-r1. I've tried unsetting all the clever-looking optimisations in |
| 19 |
the kernel, I've tried all three scheduling algorithms and I've tried |
| 20 |
resetting the BIOS to "optimised" defaults. I've even tried a genkernel |
| 21 |
kernel with default config, but that version couldn't see the root |
| 22 |
disk /dev/sda for some reason, and of course it wouldn't boot. |
| 23 |
|
| 24 |
It's also odd that CPU1 runs 5 - 6 C hotter than CPU0, whether loaded or |
| 25 |
not. |
| 26 |
|
| 27 |
Sometimes I suspect a problem with APIC or perhaps the IOMMU, re which I |
| 28 |
have mostly default or conservative settings in the kernel. Has anyone here |
| 29 |
some experience they could offer? |
| 30 |
|
| 31 |
I've also been to the BOINC project sites and changed my preferences to the |
| 32 |
most conservative I can find, but still I can't get proper allocation of |
| 33 |
boinc clients to processors. I've tried the forums and got some useful |
| 34 |
help, but not yet a solution. |
| 35 |
|
| 36 |
This all started some time ago, about the time when I had to replace the |
| 37 |
motherboard, but as I wasn't following it very closely at the time I |
| 38 |
haven't been able to pinpoint the factor that caused the change in kernel |
| 39 |
scheduling behaviour. |
| 40 |
|
| 41 |
-- |
| 42 |
Rgds |
| 43 |
Peter Humphrey |
| 44 |
Linux Counter 5290, Aug 93 |
| 45 |
-- |
| 46 |
gentoo-amd64@g.o mailing list |
Archives