| 1 |
On 11/28/2010 08:53 PM, App Deb wrote: |
| 2 |
> On Sun, Nov 28, 2010 at 7:39 PM, walt<w41ter@×××××.com> wrote: |
| 3 |
> |
| 4 |
>> I'm confused about which of all these various mechanisms apply to single-cpu |
| 5 |
>> machines. AFAICT Con's BFS (e.g.) is really a CPU scheduler and doesn't |
| 6 |
>> affect |
| 7 |
>> single-cpu machines very much. What about CFQ and group scheduling? |
| 8 |
>> Others? |
| 9 |
>> |
| 10 |
>> Thanks for any clues. |
| 11 |
> |
| 12 |
> Don't mix them, |
| 13 |
> |
| 14 |
> CFS --> upstream official CPU scheduler (also supports cgroups, that |
| 15 |
> got used in the 200line patch, which is useless imo) |
| 16 |
> CFQ --> upstream official I/O (disk) scheduler (afaik the only one |
| 17 |
> that supports "ionice") |
| 18 |
> |
| 19 |
> |
| 20 |
> BFS --> ck's CPU scheduler |
| 21 |
> (I don't know what i/o scheduler ck's patchset uses) |
| 22 |
|
| 23 |
The i/o scheduler doesn't change. |
Archives