Messages in this thread | | | Subject | Re: [git] CFS-devel, group scheduler, fixes | From | Mike Galbraith <> | Date | Wed, 19 Sep 2007 09:51:01 +0200 |
| |
On Wed, 2007-09-19 at 08:28 +0200, Mike Galbraith wrote: > Greetings, > > On Tue, 2007-09-18 at 23:03 -0700, Tong Li wrote: > > This patch attempts to improve CFS's SMP global fairness based on the new > > virtual time design. > > > > Removed vruntime adjustment in set_task_cpu() as it skews global fairness. > > Since I'm (still) encountering Xorg latency issues (which go away if > load is hefty instead of light) even with that migration adjustment and > synchronization, and am having difficulty nailing it down to a specific > event, I'll test this immediately.
(had to apply manually to freshly pulled tree)
Drat. This didn't cure the latency hits with a Xorg at nice -5 running with a make -j2 at nice 0, but seems to have reinstated a latency issue which was previously cured.
Xorg 1 sec. max latency samples: (trimmed to only show >20ms latencies) se.wait_max : 23343582 se.wait_max : 20119460 se.wait_max : 20771573 se.wait_max : 21084567 se.wait_max : 31338500 se.wait_max : 35368148 se.wait_max : 39199642 se.wait_max : 22889062 se.wait_max : 40285501 se.wait_max : 21002720 se.wait_max : 21002266 se.wait_max : 21680578 se.wait_max : 22012913 se.wait_max : 94646331 se.wait_max : 29003693 se.wait_max : 20812613
(boot with maxcpus=1 or nail X+make to one cpu and these latencies are gone, so it does seem to be the migration logic - why i was so interested in testing your patch)
The scenario which was previously cured was this: taskset -c 1 nice -n 0 ./massive_intr 2 9999 taskset -c 1 nice -n 5 ./massive_intr 2 9999 click link (http://pages.cs.wisc.edu/~shubu/talks/cachescrub-prdc2004.ppt) to bring up browser and OpenOffice Impress.
Xorg (at nice -5 + above scenario) latency samples: se.wait_max : 57985337 se.wait_max : 25163510 se.wait_max : 37005538 se.wait_max : 66986511 se.wait_max : 53990868 se.wait_max : 80976761 se.wait_max : 96967501 se.wait_max : 80989254 se.wait_max : 53990897 se.wait_max : 181963905 se.wait_max : 85985181
-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/
| |