Messages in this thread | ![/](/images/icornerl.gif) | | Date | Sat, 21 Feb 2015 20:15:27 +0100 | From | Borislav Petkov <> | Subject | Re: [RFC PATCH] x86, fpu: Use eagerfpu by default on all CPUs |
| |
On Sat, Feb 21, 2015 at 07:39:52PM +0100, Ingo Molnar wrote: > So the workload improved by ~600,000 usecs, and there's > 68,000 less calls, so it saved 8.8 usecs per call. Isn't
I think you mean more calls. The eager measurement has more calls. Let me do some primitive math:
def =(234.681331200 / 712000)*10^6 = 329.60861123595505000000 microsecs/call eager=(234.066525648 / 780000)*10^6 = 300.08528929230769000000 microsecs/call
diff is 29.52332194364736000000 microsecs speedup per call which could explain the cost of CR0.TS serialization semantics in the lazy mode.
> that a bit too high?
Now, is 29 microseconds too high? I'm not sure this is even correct and not some noise interfering.
> I'd sleep a lot better if we had some runtime debug flag to > be able to do run-to-run comparisons on the same booted up > kernel, or so.
Let me take a look whether we could so some knob... The nice thing is, code uses use_eager_fpu() to check stuff so we should be able to clear the cpufeature flag.
-- Regards/Gruss, Boris.
ECO tip #101: Trim your mails when you reply. --
| ![\](/images/icornerr.gif) |