Re: Schedulers benchmark - Was: [ANNOUNCE][RFC] PlugSched-5.2.4 for 2.6.12 and 2.6.13-rc6

[Con Kolivas <kernel@kolivas.org>]

[-- Attachment #1: Type: text/plain, Size: 496 bytes --]

On Thu, 18 Aug 2005 04:04, Michal Piotrowski wrote:
> Hi,
> here are additional staircase scheduler benchmarks.
>
> (make all -j8)
>
> scheduler:
> staircase
>
> sched_compute=1

> real    49m48.619s
> user    77m20.788s
> sys     6m7.653s

Very nice thank you.

Since you are benchmarking, here is an unofficial update to the staircase 
patch to the current non-plugsched version which will affect the results of 
interbench (it won't change the benchmark results of compute mode).

Cheers,
Con

[-- Attachment #2: staircase_update.patch --]
[-- Type: text/x-diff, Size: 751 bytes --]

Index: linux-2.6.13-rc6-plugsched/kernel/staircase.c
===================================================================
--- linux-2.6.13-rc6-plugsched.orig/kernel/staircase.c	2005-08-18 07:32:34.000000000 +1000
+++ linux-2.6.13-rc6-plugsched/kernel/staircase.c	2005-08-18 07:33:10.000000000 +1000
@@ -64,8 +64,8 @@ int sched_compute = 0;
  *compute setting is reserved for dedicated computational scheduling
  *and has ten times larger intervals.
  */
-#define _RR_INTERVAL		((10 * HZ / 1000) ? : 1)
-#define RR_INTERVAL()		(_RR_INTERVAL * (1 + 9 * sched_compute))
+#define _RR_INTERVAL		((5 * HZ / 1000) ? : 1)
+#define RR_INTERVAL()		(_RR_INTERVAL * (1 + 19 * sched_compute))
 
 #define TASK_PREEMPTS_CURR(p, rq) \
 	((p)->prio < (rq)->curr->prio)