devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Rename pones -> ptrue. Use _CMP_TRUE_UQ where appropriate.

Browse Source
This commit is contained in:
Rasmus Munk Larsen
2019-01-09 17:20:33 -08:00
parent a05ec7993e f6ba6071c5
commit fcfced13ed
9 changed files with 50 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
Eigen/src/Core/arch/AVX/PacketMath.h
View File

@@ -251,22 +251,34 @@ template<> EIGEN_STRONG_INLINE Packet8f pfloor<Packet8f>(const Packet8f& a) { re
template<> EIGEN_STRONG_INLINE Packet4d pfloor<Packet4d>(const Packet4d& a) { return _mm256_floor_pd(a); }
template<> EIGEN_STRONG_INLINE Packet8i ptrue<Packet8i>(const Packet8i& a) {
#ifdef EIGEN_VECTORIZE_AVX2
template<> EIGEN_STRONG_INLINE Packet8i pones<Packet8i>(const Packet8i& a) {
return _mm256_cmpeq_epi64(a,a);
}
// vpcmpeqd has lower latency than the more general vcmpps
return _mm256_cmpeq_epi32(a,a);
#else
template<> EIGEN_STRONG_INLINE Packet8i pones<Packet8i>(const Packet8i& /*a*/) {
const unsigned int o = 0xffffffffu;
return _mm256_set_epi32(o, o, o, o, o, o, o, o);
}
const __m256 b = _mm256_castsi256_ps(a);
return _mm256_castps_si256(_mm256_cmp_ps(b,b,_CMP_TRUE_UQ));
#endif
template<> EIGEN_STRONG_INLINE Packet8f pones<Packet8f>(const Packet8f& a) {
return _mm256_castsi256_ps(pones<Packet8i>(_mm256_castps_si256(a)));
}
template<> EIGEN_STRONG_INLINE Packet4d pones<Packet4d>(const Packet4d& a) {
return _mm256_castsi256_pd(pones<Packet8i>(_mm256_castpd_si256(a)));
template<> EIGEN_STRONG_INLINE Packet8f ptrue<Packet8f>(const Packet8f& a) {
#ifdef EIGEN_VECTORIZE_AVX2
// vpcmpeqd has lower latency than the more general vcmpps
const __m256i b = _mm256_castps_si256(a);
return _mm256_castsi256_ps(_mm256_cmpeq_epi32(b,b));
#else
return _mm256_cmp_ps(a,a,_CMP_TRUE_UQ);
#endif
}
template<> EIGEN_STRONG_INLINE Packet4d ptrue<Packet4d>(const Packet4d& a) {
#ifdef EIGEN_VECTORIZE_AVX2
// vpcmpeqq has lower latency than the more general vcmppd
const __m256i b = _mm256_castpd_si256(a);
return _mm256_castsi256_pd(_mm256_cmpeq_epi64(b,b));
#else
return _mm256_cmp_pd(a,a,_CMP_TRUE_UQ);
#endif
}
template<> EIGEN_STRONG_INLINE Packet8f pand<Packet8f>(const Packet8f& a, const Packet8f& b) { return _mm256_and_ps(a,b); }