Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-10 Thread Peter Zijlstra
On Wed, 2012-10-10 at 13:29 +0100, Mel Gorman wrote: > Do we really switch more though? > > Look at the difference in interrupts vs context switch. IPIs are an interrupt > so if TTWU_QUEUE wakes process B using an IPI, does that count as a context > switch? Nope. Nor would it for NO_TTWU_QUEUE. A

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-10 Thread Mike Galbraith
On Wed, 2012-10-10 at 13:29 +0100, Mel Gorman wrote: > On Wed, Oct 03, 2012 at 03:30:01PM +0200, Mike Galbraith wrote: > > nohz=off, pipe-test with one half pinned to CPU0, the other to CPU1. > > > > procs ---memory-- ---swap-- -io -system-- > > -cpu-- > > r b

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-10 Thread Mel Gorman
On Wed, Oct 03, 2012 at 03:30:01PM +0200, Mike Galbraith wrote: > On Wed, 2012-10-03 at 10:13 +0200, Mike Galbraith wrote: > > > Watching all cores instead. > > > > switch rate ~890KHzswitch rate ~570KHz > > NO_TTWU_QUEUE nohz=off TTWU_QUEUE no

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-05 Thread Mel Gorman
On Wed, Oct 03, 2012 at 11:04:16AM -0700, Rick Jones wrote: > On 10/03/2012 02:47 AM, Mel Gorman wrote: > >On Tue, Oct 02, 2012 at 03:48:57PM -0700, Rick Jones wrote: > >>On 10/02/2012 01:45 AM, Mel Gorman wrote: > >> > >>>SIZE=64 > >>>taskset -c 0 netserver > >>>taskset -c 1 netperf -t UDP_STREAM

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-03 Thread Rick Jones
On 10/03/2012 02:47 AM, Mel Gorman wrote: On Tue, Oct 02, 2012 at 03:48:57PM -0700, Rick Jones wrote: On 10/02/2012 01:45 AM, Mel Gorman wrote: SIZE=64 taskset -c 0 netserver taskset -c 1 netperf -t UDP_STREAM -i 50,6 -I 99,1 -l 20 -H 127.0.0.1 -- -P 15895 -s 32768 -S 32768 -m $SIZE -M $SIZE

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-03 Thread Mike Galbraith
On Wed, 2012-10-03 at 10:13 +0200, Mike Galbraith wrote: > Watching all cores instead. > > switch rate ~890KHzswitch rate ~570KHz > NO_TTWU_QUEUE nohz=off TTWU_QUEUE nohz=off > 5.38% [kernel] [k] __schedule4.81% [kernel] [k

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-03 Thread Eric Dumazet
On Wed, 2012-10-03 at 10:47 +0100, Mel Gorman wrote: > On Tue, Oct 02, 2012 at 03:48:57PM -0700, Rick Jones wrote: > > PS - I trust it is the receive-side throughput being reported/used > > with UDP_STREAM :) > > Good question. Now that I examine the scripts, it is in fact the sending > side that

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-03 Thread Mel Gorman
On Tue, Oct 02, 2012 at 03:48:57PM -0700, Rick Jones wrote: > On 10/02/2012 01:45 AM, Mel Gorman wrote: > > >SIZE=64 > >taskset -c 0 netserver > >taskset -c 1 netperf -t UDP_STREAM -i 50,6 -I 99,1 -l 20 -H 127.0.0.1 -- -P > >15895 -s 32768 -S 32768 -m $SIZE -M $SIZE > > Just FYI, unless you are

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-03 Thread Mike Galbraith
On Wed, 2012-10-03 at 08:50 +0200, Mike Galbraith wrote: > On Tue, 2012-10-02 at 14:14 +0100, Mel Gorman wrote: > > On Tue, Oct 02, 2012 at 11:31:22AM +0200, Mike Galbraith wrote: > > > On Tue, 2012-10-02 at 09:45 +0100, Mel Gorman wrote: > > > > On Tue, Oct 02, 2012 at 09:49:36AM +0200, Mike Ga

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Mike Galbraith
On Tue, 2012-10-02 at 14:14 +0100, Mel Gorman wrote: > On Tue, Oct 02, 2012 at 11:31:22AM +0200, Mike Galbraith wrote: > > On Tue, 2012-10-02 at 09:45 +0100, Mel Gorman wrote: > > > On Tue, Oct 02, 2012 at 09:49:36AM +0200, Mike Galbraith wrote: > > > > > > Hm, 518cd623 fixed up the troubles I s

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Rick Jones
On 10/02/2012 01:45 AM, Mel Gorman wrote: SIZE=64 taskset -c 0 netserver taskset -c 1 netperf -t UDP_STREAM -i 50,6 -I 99,1 -l 20 -H 127.0.0.1 -- -P 15895 -s 32768 -S 32768 -m $SIZE -M $SIZE Just FYI, unless you are running a hacked version of netperf, the "50" in "-i 50,6" will be silently

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Mike Galbraith
On Tue, 2012-10-02 at 14:14 +0100, Mel Gorman wrote: > On Tue, Oct 02, 2012 at 11:31:22AM +0200, Mike Galbraith wrote: > > On Tue, 2012-10-02 at 09:45 +0100, Mel Gorman wrote: > > > On Tue, Oct 02, 2012 at 09:49:36AM +0200, Mike Galbraith wrote: > > > > > > Hm, 518cd623 fixed up the troubles I s

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Mel Gorman
On Tue, Oct 02, 2012 at 11:31:22AM +0200, Mike Galbraith wrote: > On Tue, 2012-10-02 at 09:45 +0100, Mel Gorman wrote: > > On Tue, Oct 02, 2012 at 09:49:36AM +0200, Mike Galbraith wrote: > > > > Hm, 518cd623 fixed up the troubles I saw. How exactly are you running > > > this? > > > > > > > You

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Mike Galbraith
On Tue, 2012-10-02 at 09:45 +0100, Mel Gorman wrote: > On Tue, Oct 02, 2012 at 09:49:36AM +0200, Mike Galbraith wrote: > > Hm, 518cd623 fixed up the troubles I saw. How exactly are you running > > this? > > > > You saw problems with TCP_RR where as this is UDP_STREAM. Yeah, but I wanted to st

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Mel Gorman
On Tue, Oct 02, 2012 at 09:49:36AM +0200, Mike Galbraith wrote: > On Tue, 2012-10-02 at 07:51 +0100, Mel Gorman wrote: > > I'm going through old test results to see could I find any leftover > > performance regressions that have not yet been fixed (most have at this > > point > > or at least chan

Re: Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-02 Thread Mike Galbraith
On Tue, 2012-10-02 at 07:51 +0100, Mel Gorman wrote: > I'm going through old test results to see could I find any leftover > performance regressions that have not yet been fixed (most have at this point > or at least changed in such a way to make a plain revert impossible). One > major regression

Netperf UDP_STREAM regression due to not sending IPIs in ttwu_queue()

2012-10-01 Thread Mel Gorman
I'm going through old test results to see could I find any leftover performance regressions that have not yet been fixed (most have at this point or at least changed in such a way to make a plain revert impossible). One major regression still left is with netperf UDP_STREAM regression. Bisection po