Messages in this thread | | | Subject | Re: volanoMark regression with kernel 2.6.26-rc1 | From | Mike Galbraith <> | Date | Wed, 14 May 2008 16:50:33 +0200 |
| |
On Wed, 2008-05-14 at 19:14 +0530, Srivatsa Vaddagiri wrote: > On Mon, May 12, 2008 at 11:20:47AM +0200, Peter Zijlstra wrote: > > It looks like things like find_busiest_group() just think everything is > > peachy when the imbalance is < 1 task - which with all this grouping > > stuff is not necessarily true. > > fwiw, the following hack seems to help bring down regression to ~5% > (b/n 2.6.25 and 2.6.26-rc1 with USER_SCHED): > > Also in the patch GROUP_SCALE can probably be less than what I used (and > needs a ifdef GROUP_SCHED) ..
On my little Q6600 with git.today, it made no real difference, whether NORMALIZED_SLEEPERS was enabled or not. I see roughly 10-15% idle time w/wo this patch, whereas pre-regression, it's < 2%.
-Mike
| |