Re: [RFC PATCH] x86, fpu: Use eagerfpu by default on all CPUs

[Borislav Petkov]

On Sun, Feb 22, 2015 at 09:18:40AM +0100, Ingo Molnar wrote:
>   - It might make sense to do a 'perf stat --null --repeat'
>     measurement as well [without any -e arguments], to make 
>     sure the rich PMU stats you are gathering are not 
>     interfering?

Well, the --repeat thing definitely is good to do and the previous
single run is simply not meaningful due to variance I'd guess:

perf stat --null --repeat 10 --sync --pre ~/bin/pre-build-kernel.sh make -s -j12

...

Lazy FPU:
  FPU saves before: 5
Setup is 16252 bytes (padded to 16384 bytes).
System is 4222 kB
CRC 5417fa4f
Kernel: arch/x86/boot/bzImage is ready  (#43)

...

Kernel: arch/x86/boot/bzImage is ready  (#52)

 Performance counter stats for 'make -s -j12' (10 runs):

     219.406449195 seconds time elapsed                                          ( +-  0.17% )

  FPU saves after: 6974975

Eager FPU:
  FPU saves before: 4
Setup is 16252 bytes (padded to 16384 bytes).
System is 4222 kB
CRC 823c1801
Kernel: arch/x86/boot/bzImage is ready  (#53)

...

Kernel: arch/x86/boot/bzImage is ready  (#62)

 Performance counter stats for 'make -s -j12' (10 runs):

     218.791122148 seconds time elapsed                                          ( +-  0.13% )

  FPU saves after: 8939084


The counter numbers are consistent with before: ~200000 FPU saves more
in eager mode per kernel build.

Timing improvement of 0.6 secs on average looks nice to me. Next I'll
do the same thing in tmpfs to see how the numbers look without the page
cache.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/