Re: Spinlocks, yet again: analysis and proposed patches

Josh Berkus <josh(at)agliodbs(dot)com> writes:

> Tom,
>
> > What I think this means is that the kernel is scheduling the 2 processes
> > onto 2 processors chosen-at-random, without awareness of whether those
> > two processors are on the same chip (in the Xeon case) or have closer
> > NUMA affinity (in the Opteron case).
>
> That would be consistent with my experience with HT, and the reason why many
> software vendors recommend disabling it. Not sure about NUMA.

What version of linux was this test with? What you describe was certainly well
known for Linux 2.4 and earlier. It was well on its way to becoming
established dogma cargo-cult style.

However I was under the impression that 2.6 had moved beyond that problem.
It would be very interesting to know if 2.6 still suffers from this.

Also, those who recommend disabling HT on this basis should realize it didn't
penalize the >2 process case the same way. You may be getting a benefit only
in the cases where your system isn't heavily loaded. Ie, only when it doesn't
really matter.

--
greg