... a parallel life due to Subject: changes, dropped Cc:'s, and ... improve the numbers? > > Hmm, I know diddly about Java, and I ... Each thread runs multiple queries > > with a 10ms pause (to allow ... table also need to implement some form of locking? How would ...

linux-kernel - Satyam Sharma - Sep 14 2007 - 12:01

... > does this improve the numbers? Hmm, I know diddly about Java, and I don't ... loop. Each thread runs multiple queries > with a 10ms pause (to allow the ... cpus (2) > * more general information (which also covers some irrelevant > information about some ...

linux-kernel - Satyam Sharma - Sep 14 2007 - 06:06

... (see above) so I'll re-test -rc3 first, git-bisect ... between tests takes about 30mins. >> * java threads are created first and ... Don't know much about CFS either, but does that constant ... lead to different numbers? I've tested before with varying timings ...

linux-kernel - Antoine Martin - Sep 14 2007 - 11:25

... : maybe, the locality of the java threads is important, but there are ... "pathological". And CFS, does show some noticeable improvements with the new measurements ... Kernels-ManyThreads-CombinedTests8-5msYield.png The number of threads that exit before we ...

linux-kernel - Antoine Martin - Sep 17 2007 - 08:17

... 2.6 we changed it a number of times. To avoid the ... , 88 insertions(+), 6 deletions(-) Ingo ------------------------------> Subject: sched: yield workaround From: Ingo Molnar ... fundamentally broken, and CFS has changed its behavior. Some apps that ...

linux-kernel - Ingo Molnar - Sep 19 2007 - 15:18

... default. > > I disagree. I think CFS made "sched_yield()" worse, and what ... and quite frankly, the current CFS behaviour simply looks buggy. It ... fully emulate it in CFS. in CFS we dont have a ... apps that do calls to yield() [some firefox plugins do that ...

linux-kernel - Ingo Molnar - Sep 19 2007 - 15:56

... you get it to work fast with the -rc6+yield-patch solution too ... anymore but always available. Ingo -------------------> Subject: sched: yield workaround From: Ingo Molnar ... is fundamentally broken, and CFS has changed its behavior. Some apps that ...

linux-kernel - Ingo Molnar - Sep 14 2007 - 11:32

... =295071 ================================================== Description of problem: The CFS scheduler does not seem to ... .22.5-76. Version-Release number of selected component (if applicable ... that it is not some anomaly with the way top calculates ...

linux-kernel - Chuck Ebbert - Sep 18 2007 - 19:02

... timeslice (or whatever the equivalent is in CFS) until the > yielding task again "becomes head ... this isn't happening? I'll run some tests on my SMP system running CFS. But I'll ... what else could be happening here. Does CFS spin in a loop doing nothing when ...

linux-kernel - David Schwartz - Sep 19 2007 - 18:56

... on by default. I disagree. I think CFS made "sched_yield()" worse, and what you call ... scheduler queue" like you do now in CFS. It very much is about moving the ... to run. and quite frankly, the current CFS behaviour simply looks buggy. It should ...

linux-kernel - Linus Torvalds - Sep 19 2007 - 15:39

... to fully emulate it in > CFS. > > in CFS we dont have a per- ... desktop apps > that do calls to yield() [some firefox plugins do that, ... - but this would be expected.) Ingo ----------------------> Subject: sched: make yield more agressive From: ...

linux-kernel - Ingo Molnar - Sep 19 2007 - 16:26

... might complain about _that_. Yeah, with reasonably good reason. The sched_yield ... think we have two problems with yield(): - it really doesn't ... what CFS does now. And if we then have some agreement ... a good thing - maybe we're not perfect, but at least ...

linux-kernel - Linus Torvalds - Sep 19 2007 - 16:28

... this improve the numbers? the patch i sent was against CFS-devel. ... -rc6, does it improve the numbers? (it should have an impact) ... able to twiddle the sysctl. Ingo ------------------------> Subject: sched: yield debugging From: Ingo Molnar ...

linux-kernel - Ingo Molnar - Sep 14 2007 - 04:32

... /Linux-Kernels/Kernels-ManyThreads-CombinedTests5-10msYield.png >> Compared with more kernels here - a bit more cluttered: >> ... - can you get it to work fast with the > -rc6+yield-patch solution too? (i.e. not CFS-devel) We need a (tested) > solution ...

linux-kernel - Chuck Ebbert - Sep 18 2007 - 13:00

... Kernels/Kernels-ManyThreads-CombinedTests5-10msYield.png > >> Compared with more kernels here - a bit more ... can you get it to work fast with the > > -rc6+yield-patch solution too? (i.e. not CFS-devel) ... for 2.6.23 and the CFS-devel patches are not for 2.6 ...

linux-kernel - Ingo Molnar - Sep 18 2007 - 18:46

... /task2" This seems right. They're both always ready to run. ... that it is not some > anomaly with the > way top calculates ... its thread list. It takes no arguments." CFS is perfectly complying ... punished). (Not that I agree with this argument, I'm just ...

linux-kernel - David Schwartz - Sep 19 2007 - 14:45

... timeslice (or whatever the equivalent is in CFS) until the > > yielding task again "becomes ... this isn't happening? I'll run some tests on my SMP > system running CFS. But I'll ... task should get less CPU since they're both always ready-to-run. The only ...

linux-kernel - David Schwartz - Sep 19 2007 - 19:05

... s possible that we could perhaps implement > > the CFS "yield()" using the same logic as we ... ) or the second (if they > want some massive delay). So while it does not ... me a little. Also, while I agree with Linus' definition of sched_yield, I'm afraid ...

linux-kernel - Peter Zijlstra - Sep 19 2007 - 17:58

... task has given up the cpu. The other task should get to run for a timeslice (or whatever the equivalent is in CFS) until the yielding task again "becomes head of the thread list". Chris - To unsubscribe from this list: send the line "unsubscribe ...

linux-kernel - Chris Friesen - Sep 19 2007 - 15:48

... . (The only well-defined behavior of yield > is for SCHED_FIFO/RR tasks - and that is fully preserved in CFS.) Certainly this is reasonable for applications for which the source is available and readily recompilable. However, there are legacy ...

linux-kernel - Chris Friesen - Sep 19 2007 - 16:00