[haiku-development] Re: Optimizing Painter::_DrawBitmapBilinearCopy32
- From: Urias McCullough <umccullough@xxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Sun, 14 Jun 2009 17:31:27 +0000
On Sun, Jun 14, 2009 at 5:04 PM, Urias McCullough<umccullough@xxxxxxxxx> wrote:
> Note, this machine is running gcc4 Haiku r30980 - which I suspect
> doesn't matter really.
>
> ~> runme_haiku
> Benchmark: Haiku app_server bilinear copy
> Compile date: Jun 14 2009 14:38:02
> GCC version: 2.95.3-haiku-081024
>
> CPU vendor ID: GenuineIntel
> CPU: Intel(R) Atom(TM) CPU N270 @ 1.60GHz
> SIMD instructions: MMX SSE SSE-Integer SSE2 SSE3 SSSE3
>
> Can't lock process to CPU on this platform.
> Estimated CPUID/RDTSC overhead: 132 clock cycles.
> 10 runs per benchmark.
>
> -- Results --
>
> Minimum Average Maximum
> # 1: 1046784 1074346 1282392 - 'C, original'
> # 2: 1164516 1173552 1206912 - 'C, precise'
> # 3: 1288284 1293818 1305444 - 'C, precise DIV'
> # 4: 520128 521977 530472 - 'MMX/SSE'
> # 5: 454704 461757 488664 - 'MMX/SSE optim-test'
> # 6: 482016 483688 492480 - 'SSE2'
> # 7: 515856 517687 526404 - 'SSSE3'
Running it on a gcc2 Haiku on the same machine gave slightly lower max
values on the "C" code, which is strange since i didn't think it was
using any of the OS functions?
/Haiku1/home> runme_haiku
Benchmark: Haiku app_server bilinear copy
Compile date: Jun 14 2009 14:38:02
GCC version: 2.95.3-haiku-081024
CPU vendor ID: GenuineIntel
CPU: Intel(R) Atom(TM) CPU N270 @ 1.60GHz
SIMD instructions: MMX SSE SSE-Integer SSE2 SSE3 SSSE3
Can't lock process to CPU on this platform.
Estimated CPUID/RDTSC overhead: 348 clock cycles.
10 runs per benchmark.
-- Results --
Minimum Average Maximum
# 1: 1046472 1062784 1187568 - 'C, original'
# 2: 1164288 1167100 1174560 - 'C, precise'
# 3: 1287912 1302430 1350480 - 'C, precise DIV'
# 4: 519972 521715 530304 - 'MMX/SSE'
# 5: 454488 457060 479532 - 'MMX/SSE optim-test'
# 6: 481896 483454 491520 - 'SSE2'
# 7: 515664 517351 525516 - 'SSSE3'
Other related posts:
