> On Feb 4, 2020, at 12:39 PM, Waiman Long <longman@xxxxxxxxxx> wrote: > > On 2/4/20 12:27 PM, Peter Zijlstra wrote: >> On Tue, Feb 04, 2020 at 11:54:02AM -0500, Alex Kogan wrote: >>>> On Feb 3, 2020, at 10:47 AM, Waiman Long <longman@xxxxxxxxxx> wrote: >>>> >>>> On 2/3/20 10:28 AM, Peter Zijlstra wrote: >>>>> On Mon, Feb 03, 2020 at 09:59:12AM -0500, Waiman Long wrote: >>>>>> On 2/3/20 8:45 AM, Peter Zijlstra wrote: >>>>>>> Presumably you have a workload where CNA is actually a win? That is, >>>>>>> what inspired you to go down this road? Which actual kernel lock is so >>>>>>> contended on NUMA machines that we need to do this? >>> There are quite a few actually. files_struct.file_lock, file_lock_context.flc_lock >>> and lockref.lock are some concrete examples that get very hot in will-it-scale >>> benchmarks. >> Right, that's all a variant of banging on the same resources across >> nodes. I'm not sure there's anything fundamental we can fix there. Not much, except gain that 2x from a better lock. >> >>> And then there are spinlocks in __futex_data.queues, >>> which get hot when applications have contended (pthread) locks — >>> LevelDB is an example. >> A numa aware rework of futexes has been on the todo list for years :/ > Now, we are going to get that for free with this patchset:-) Exactly!! >> >>> Our initial motivation was based on an observation that kernel qspinlock is not >>> NUMA-aware. So what, you may ask. Much like people realized in the past that >>> global spinning is bad for performance, and they switched from ticket lock to >>> locks with local spinning (e.g., MCS), I think everyone would agree these days that >>> bouncing a lock (and cache lines in general) across numa nodes is similarly bad. >>> And as CNA demonstrates, we are easily leaving 2-3x speedups on the table by >>> doing just that with the current qspinlock. >> Actual benchmarks with performance numbers are required. It helps >> motivate the patches as well as gives reviewers clues on how to >> reproduce / inspect the claims made. >> > I think the cover-letter does have some benchmark results listed. To clarify on that, I _used to include benchmark results in the cover letter for previous revisions. I stopped doing that as the changes between revisions were rather minor — maybe that is the missing part? If so, my apologies, I can certainly publish them again. The point is that we have numbers for actual benchmarks, plus the kernel build robot has sent quite a few reports on positive improvements in the performance of AIM7 and other benchmarks due to CNA (plus ARM folks noticed improvement in their benchmarks, although I think those were mostly microbenchmarks. Yet, it is evident that the improvements are cross-platform.) Regards, — Alex
