mirror/eigen
2
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2024-12-15 07:10:37 +08:00
Code Issues Packages Projects Releases Wiki Activity

half implementation has been moved to half_impl namespace

Browse Source
This commit is contained in:
Benoit Steiner 2016-07-29 15:09:34 -07:00
parent 2693fd54bf
commit 02fe89f5ef
1 changed files with 17 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
Eigen/src/Core/arch/CUDA/PacketMathHalf.h
View File

@ -607,7 +607,7 @@ template<> EIGEN_STRONG_INLINE Packet4h pset1<Packet4h>(const Eigen::half& from)
} }
template<> EIGEN_STRONG_INLINE Eigen::half pfirst<Packet4h>(const Packet4h& from) { template<> EIGEN_STRONG_INLINE Eigen::half pfirst<Packet4h>(const Packet4h& from) {
return raw_uint16_to_half(static_cast<unsigned short>(_mm_cvtsi64_si32(from.x))); return half_impl::raw_uint16_to_half(static_cast<unsigned short>(_mm_cvtsi64_si32(from.x)));
} }
template<> EIGEN_STRONG_INLINE Packet4h pconj(const Packet4h& a) { return a; } template<> EIGEN_STRONG_INLINE Packet4h pconj(const Packet4h& a) { return a; }
@ -618,17 +618,17 @@ template<> EIGEN_STRONG_INLINE Packet4h padd<Packet4h>(const Packet4h& a, const
Eigen::half h[4]; Eigen::half h[4];
Eigen::half ha = raw_uint16_to_half(static_cast<unsigned short>(a64)); Eigen::half ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64));
Eigen::half hb = raw_uint16_to_half(static_cast<unsigned short>(b64)); Eigen::half hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64));
h[0] = ha + hb; h[0] = ha + hb;
ha = raw_uint16_to_half(static_cast<unsigned short>(a64 >> 16)); ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64 >> 16));
hb = raw_uint16_to_half(static_cast<unsigned short>(b64 >> 16)); hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64 >> 16));
h[1] = ha + hb; h[1] = ha + hb;
ha = raw_uint16_to_half(static_cast<unsigned short>(a64 >> 32)); ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64 >> 32));
hb = raw_uint16_to_half(static_cast<unsigned short>(b64 >> 32)); hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64 >> 32));
h[2] = ha + hb; h[2] = ha + hb;
ha = raw_uint16_to_half(static_cast<unsigned short>(a64 >> 48)); ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64 >> 48));
hb = raw_uint16_to_half(static_cast<unsigned short>(b64 >> 48)); hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64 >> 48));
h[3] = ha + hb; h[3] = ha + hb;
Packet4h result; Packet4h result;
result.x = _mm_set_pi16(h[3].x, h[2].x, h[1].x, h[0].x); result.x = _mm_set_pi16(h[3].x, h[2].x, h[1].x, h[0].x);
@ -641,17 +641,17 @@ template<> EIGEN_STRONG_INLINE Packet4h pmul<Packet4h>(const Packet4h& a, const
Eigen::half h[4]; Eigen::half h[4];
Eigen::half ha = raw_uint16_to_half(static_cast<unsigned short>(a64)); Eigen::half ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64));
Eigen::half hb = raw_uint16_to_half(static_cast<unsigned short>(b64)); Eigen::half hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64));
h[0] = ha * hb; h[0] = ha * hb;
ha = raw_uint16_to_half(static_cast<unsigned short>(a64 >> 16)); ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64 >> 16));
hb = raw_uint16_to_half(static_cast<unsigned short>(b64 >> 16)); hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64 >> 16));
h[1] = ha * hb; h[1] = ha * hb;
ha = raw_uint16_to_half(static_cast<unsigned short>(a64 >> 32)); ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64 >> 32));
hb = raw_uint16_to_half(static_cast<unsigned short>(b64 >> 32)); hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64 >> 32));
h[2] = ha * hb; h[2] = ha * hb;
ha = raw_uint16_to_half(static_cast<unsigned short>(a64 >> 48)); ha = half_impl::raw_uint16_to_half(static_cast<unsigned short>(a64 >> 48));
hb = raw_uint16_to_half(static_cast<unsigned short>(b64 >> 48)); hb = half_impl::raw_uint16_to_half(static_cast<unsigned short>(b64 >> 48));
h[3] = ha * hb; h[3] = ha * hb;
Packet4h result; Packet4h result;
result.x = _mm_set_pi16(h[3].x, h[2].x, h[1].x, h[0].x); result.x = _mm_set_pi16(h[3].x, h[2].x, h[1].x, h[0].x);