本帖最後由 vsp 於 2017-2-12 10:23 編輯
AVX is expensive in die size and power consumption
qcmadness 發表於 2017-2-11 22:36 
http://www.numberworld.org/y-cru ... -optimizations.html
When the user runs the dispatcher, it walks down the list of binaries in the order they are listed above and picks the first one that will run.
The following exceptions are applied:
Both SSE4.1 binaries are blacklisted for all AMD processors since they run slower than SSE3.
AVX and AVX2 are blacklisted from AMD Bulldozer processors since they are slower than SSE3 and XOP.
XOP is blacklisted for all non-AMD processors since it's unlikely anyone other than AMD will support them. Even AMD is dropping them for the Zen processor.
If the "SSE4.1 ~ Nagisa" binary is available, then "SSE4.1 ~ Ushio" is blacklisted for Core 2 processors.
AVX512 is blacklisted for non-Intel processors on Windows because the Intel Compiler refuses to generate AVX512 that isn't vendor-locked to Intel processors.
y-cruncher v0.7.1.9465 x86 x64
Pi - 250 million digits - SSE3 SSE3 SSE4.1 AVX XOP AVX2 ADX
Processor Kasumi Ushio Hina Miyu Airi Kurumi
Core 2 Quad Q6600 Intel Core 2.4 GHz 308.968 222.807 156.763
Core i7 920 Intel Nehalem 3.5 GHz 168.149 113.313 83.244 80.000
Core i7 3630QM Intel Ivy Bridge 3.2 GHz 153.950 98.872 76.648 73.044 56.130
FX-8350 AMD Piledriver 4.0 GHz 181.973 105.597 70.267 69.988 72.020 54.761
Core i7 4770K Intel Haswell 4.0 GHz 118.395 73.312 58.699 55.169 46.588 24.227
Core i7 6820HK Intel Skylake 3.2 GHz 138.380 83.017 67.255 62.939 48.250 25.059 24.877
y-cruncher v0.7.1.9465 x64
Pi - 1 billion digits SSE3 SSE4.1 AVX XOP AVX2 ADX
Processor Kasumi Ushio Hina Miyu Airi Kurumi
Core 2 Quad Q6600 Intel Core 2.4 GHz 803.675
Core i7 920 Intel Nehalem 3.5 GHz 424.546 406.073
Core i7 3630QM Intel Ivy Bridge 3.2 GHz 404.973 393.676 313.130
FX-8350 AMD Piledriver 4.0 GHz 353.717 351.583 376.275 266.824
Core i7 4770K Intel Haswell 4.0 GHz 297.314 279.921 239.519 119.836
Core i7 6820HK Intel Skylake 3.2 GHz 338.069 319.290 248.265 122.021 121.586 |