portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-09 13:40:46 +01:00
Code Releases Activity

Compile options tweak

Browse Source
This commit is contained in:
Peter Boyle 2015-05-15 12:33:18 +01:00
parent c99922b591
commit f43589369a
2 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
configure-commands
View File

@ -3,13 +3,13 @@ WD=$1
echo $WD echo $WD
case $WD in case $WD in
build-icpc-avx) build-icpc-avx)
CXX=icpc ../configure --enable-simd=AVX CXXFLAGS="-mavx -O3 -I -std=c++11" --enable-comms=none CXX=icpc ../configure --enable-simd=AVX CXXFLAGS="-mavx -O3 -std=c++11" --enable-comms=none
;; ;;
build-icpc-avx512) build-icpc-avx512)
CXX=icpc ../configure --enable-simd=AVX512 CXXFLAGS="-xCOMMON-AVX512 -O3 -I -std=c++11" --host=none --enable-comms=none CXX=icpc ../configure --enable-simd=AVX512 CXXFLAGS="-xCOMMON-AVX512 -O3 -std=c++11" --host=none --enable-comms=none
;; ;;
build-icpc-mic) build-icpc-mic)
CXX=icpc ../configure --host=none --enable-simd=AVX512 CXXFLAGS="-mmic -O3 -I -std=c++11" LDFLAGS=-mmic --enable-comms=none CXX=icpc ../configure --host=none --enable-simd=AVX512 CXXFLAGS="-mmic -O3 -std=c++11" LDFLAGS=-mmic --enable-comms=none
;; ;;
build-clang-avx) build-clang-avx)
CXX=clang++ ../configure --enable-simd=AVX CXXFLAGS="-mavx -O3 -std=c++11" --enable-comms=none CXX=clang++ ../configure --enable-simd=AVX CXXFLAGS="-mavx -O3 -std=c++11" --enable-comms=none

1
lib/simd/Grid_vComplexF.h
View File

@ -245,6 +245,7 @@ namespace Grid {
float f[8]; float f[8];
} u256; } u256;
//SWAP lanes //SWAP lanes
// FIXME .. icc complains with lib/lattice/Grid_lattice_reduction.h (49): (col. 20) warning #13211: Immediate parameter to intrinsic call too large
__m256 t0 = _mm256_permute2f128_ps(in.v, in.v, 1); __m256 t0 = _mm256_permute2f128_ps(in.v, in.v, 1);
__m256 t1 = _mm256_permute_ps(in.v , 0b11011000);//real (0,2,1,3) __m256 t1 = _mm256_permute_ps(in.v , 0b11011000);//real (0,2,1,3)
__m256 t2 = _mm256_permute_ps(t0 , 0b10001101);//imag (1,3,0,2) __m256 t2 = _mm256_permute_ps(t0 , 0b10001101);//imag (1,3,0,2)