Messages in this thread | | | Subject | Re: [PATCH] sched: high-res preemption tick | From | Mike Galbraith <> | Date | Sun, 14 Oct 2007 08:34:20 +0200 |
| |
On Sun, 2007-10-14 at 01:13 +0200, Peter Zijlstra wrote: > On Sat, 2007-10-13 at 11:17 +0200, Peter Zijlstra wrote: > > > > Ah, but HRTICK is not compatible with PREEMPT_RESTRICT, it will be > > > similar to !WAKEUP_PREEMPT. > > > > (I do plan to fix that eventually, just need to do it) > > I guess something like this ought to do, but its a tad late so I'm quite > sure :-)
2.6.23-smp-d-hrt + restrict fix patch [SUM] 0.0-300.1 sec 176 GBytes 5.03 Gbits/sec [SUM] 0.0-300.1 sec 175 GBytes 5.02 Gbits/sec [SUM] 0.0-300.1 sec 176 GBytes 5.05 Gbits/sec
Context switches are further reduced (across the board) over PREEMPT_RESTRICT, dropping from ~7-8k to ~2.5k with this test, vs ~950 for SCHED_BATCH and ~50k with this tree and no restriction. Throughput is ~96% of SCHED_BATCH, vs ~55% with no restriction. I see no interactivity regressions.
-Mike
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |