Messages in this thread | | | Date | Mon, 1 Jul 2013 09:50:46 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH] x86: Use asm-goto to implement mutex fast path on x86-64 |
| |
* Borislav Petkov <bp@alien8.de> wrote:
> On Sat, Jun 29, 2013 at 04:56:30PM -0700, Wedson Almeida Filho wrote: > > On Fri, Jun 28, 2013 at 7:09 AM, Borislav Petkov <bp@alien8.de> wrote: > > > > > Btw, do we have any perf data showing any improvements from this patch? > > > > I wrote a simple test the measures the time it takes to acquire and > > release an uncontended mutex (i.e., we always take the fast path) > > 100k times. I ran it a few times, the original code averages > > 2.743436ms, and the new code averages 2.101098ms, so it's about 23% improvement. > > Microbenchmark results tend to be misleading in such situations. Rather, > it would be much closer to reality if you traced a real workload like a > simple kernel build, for example, with and without your patch.
Not sure - the main thing we want to know is whether it gets faster. The _amount_ will depend on things like precise usage patterns, caching, etc. - but rarely does a real workload turn a win like this into a loss.
> I.e., something like > > perf stat --repeat 5 ./build-kernel.sh > > and take a look at what the perfcouters are saying in both cases.
Hm, the noise of such a workload will very likely drown out improvements that are in the cycle scale.
Thanks,
Ingo
| |