1 |
On Sun, 20 May 2007, Duncan wrote: |
2 |
|
3 |
> Date: Sun, 20 May 2007 16:14:45 +0000 (UTC) |
4 |
> From: Duncan <1i5t5.duncan@×××.net> |
5 |
> Reply-To: gentoo-amd64@l.g.o |
6 |
> To: gentoo-amd64@l.g.o |
7 |
> Subject: [gentoo-amd64] Re: Kernel scheduler seems to be making mistakes |
8 |
> |
9 |
> Peter Humphrey <prh@××××××××××.uk> posted |
10 |
> 200705201511.52195.prh@××××××××××.uk, excerpted below, on Sun, 20 May |
11 |
> 2007 15:11:52 +0100: |
12 |
> So, if you emerge schedutils, one of the binaries you get is called |
13 |
> taskset. Once it's emerged, you can read the notes on taskset in /usr/ |
14 |
> share/doc/schedutils-*/README.bz2, and/or the taskset manpage. It's |
15 |
> pretty simple, to use, however. For example, on a two-core or two-CPU |
16 |
> system (CPU0 and CPU1), setting an already running X to run on CPU0 only, |
17 |
> on CPU1 only, or on both, is done with the following commands (the number |
18 |
> being a CPU bitmask, obviously): |
19 |
> |
20 |
> taskset -p 1 `pidof X` |
21 |
> taskset -p 2 `pidof X` |
22 |
> taskset -p 3 `pidof X` |
23 |
|
24 |
The problem with that is that the boinc client starts a new thread for |
25 |
every work unit. |
26 |
|
27 |
I've noticed the same problem and I think that it should be up to boinc to |
28 |
set that up properly |
29 |
-- |
30 |
gentoo-amd64@g.o mailing list |