Re: Quitte à faire du branchless

Bien vu pour la version sans multiplication :) On voit bien la vectorisation et ça marche bien mieux :

2020年02月12日 08:34:50
Running ./build/matrix-elements-sum-GNU
Run on (4 X 3100 MHz CPU s)
CPU Caches:
 L1 Data 32 KiB (x2)
 L1 Instruction 32 KiB (x2)
 L2 Unified 256 KiB (x2)
 L3 Unified 3072 KiB (x1)
Load Average: 0.67, 0.67, 0.63
***WARNING*** CPU scaling is enabled, the benchmark real time measurements may be noisy and will incur extra overhead.
-------------------------------------------------------------
Benchmark Time CPU Iterations
-------------------------------------------------------------
branchless/2 28.4 ns 28.4 ns 24540986
branchless/4 45.6 ns 45.6 ns 15164641
branchless/8 116 ns 116 ns 5777179
branchless/16 369 ns 369 ns 1898641
branchless/32 1387 ns 1387 ns 484254
branchless/64 5847 ns 5846 ns 111630
branchless/128 22439 ns 22436 ns 30474
branchless/256 88132 ns 88125 ns 7754
branchless/512 349637 ns 349573 ns 2010
branchless/1024 1441557 ns 1441252 ns 485
branchless/2048 5766216 ns 5764464 ns 121
branchless/4096 22847843 ns 22846967 ns 31
branchless/8192 91376681 ns 91368790 ns 8
branchless/16384 367024190 ns 366997034 ns 2
branchless_2/2 34.7 ns 34.7 ns 20037235
branchless_2/4 40.5 ns 40.5 ns 17081915
branchless_2/8 62.9 ns 62.9 ns 10383991
branchless_2/16 146 ns 146 ns 4770697
branchless_2/32 435 ns 435 ns 1600480
branchless_2/64 1869 ns 1863 ns 450005
branchless_2/128 5882 ns 5882 ns 104268
branchless_2/256 24406 ns 24407 ns 28618
branchless_2/512 100493 ns 100124 ns 7202
branchless_2/1024 473645 ns 473618 ns 1248
branchless_2/2048 2284499 ns 2284170 ns 288
branchless_2/4096 6951965 ns 6951809 ns 79
branchless_2/8192 33005646 ns 33002903 ns 20
branchless_2/16384 118768214 ns 118758699 ns 5