portelli/Grid
1
0
Fork 0
mirror of https://github.com/paboyle/Grid.git synced 2025-07-31 19:57:07 +01:00
Code Releases Activity

Added missing SIMD integer reduction implementation for AVX, AVX-512, SSE4, IMCI

Browse Source
This commit is contained in:
Lanny91
2017-06-16 15:58:47 +01:00
parent 07b2c1b253
commit a833f88c32
4 changed files with 45 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
lib/simd/Grid_avx.h
View File

@@ -701,9 +701,28 @@ namespace Optimization {
//Integer Reduce
template<>
inline Integer Reduce<Integer, __m256i>::operator()(__m256i in){
// FIXME unimplemented
printf("Reduce : Missing integer implementation -> FIX\n");
assert(0);
__m128i ret;
#if defined (AVX2)
// AVX2 horizontal adds within upper and lower halves of register; use
// SSE to add upper and lower halves for result.
__m256i v1, v2;
__m128i u1, u2;
v1 = _mm256_hadd_epi32(in, in);
v2 = _mm256_hadd_epi32(v1, v1);
u1 = _mm256_castsi256_si128(v2); // upper half
u2 = _mm256_extracti128_si256(v2, 1); // lower half
ret = _mm256_add_epi32(u1, u2);
#else
// No AVX horizontal add; extract upper and lower halves of register & use
// SSE intrinsics.
__m128i u1, u2, u3;
u1 = _mm256_extractf128_si256(in, 0); // upper half
u2 = _mm256_extractf128_si256(in, 1); // lower half
u3 = _mm_add_epi32(u1, u2);
u1 = _mm_hadd_epi32(u3, u3);
ret = _mm_hadd_epi32(u1, u1);
#endif
return _mm_cvtsi128_si32(ret);
}
}