SIMD Matrix Vectorization:使用Intel AVX Intrinsics对矩阵进行向量化。 具有其他优化功能可以利用局部性来减少高速缓存
SIMD矩阵矢量化 依赖项/安装 Your CPU must support Intel AVX Intrinsics 运行代码 编译: Run the compile_all.sh bash script from the source directory with sh compile_all.sh 跑步: Run the tester_transpose.sh or tester_transpose_block.sh (or other) script from the source directory with sh tester_transpose_block.sh (This outputs a bunch of performance info in CSV format, which is available in a much more readable form
文件列表
SIMD-Matrix-Vectorization-master.zip
(预估有个36文件)
SIMD-Matrix-Vectorization-master
Part1
CSVs
outputSequential.txt
2KB
localMachineOutputBlock.txt
849B
localMachineOutputSequential.txt
2KB
outputBlock.txt
852B
Source
transpose_block_bash
15KB
transpose_bash
15KB
暂无评论