mirror/eigen
2
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-02-05 17:50:26 +08:00
Code Issues Packages Projects Releases Wiki Activity

Reverse compare logic ƒin F32ToBf16 since vec_cmpne is not available in Power8 - now compiles for clang10 default (P8).

Browse Source
This commit is contained in:
Chip-Kerchner 2021-08-13 11:21:28 -05:00
parent 66499f0f17
commit e07227c411
1 changed files with 7 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
Eigen/src/Core/arch/AltiVec/PacketMath.h
View File

@ -902,8 +902,8 @@ template<> EIGEN_STRONG_INLINE Packet8bf pxor<Packet8bf>(const Packet8bf& a, con
return pxor<Packet8us>(a, b);
}
template<> EIGEN_STRONG_INLINE Packet4f pandnot<Packet4f>(const Packet4f& a, const Packet4f& b) { return vec_and(a, vec_nor(b, b)); }
template<> EIGEN_STRONG_INLINE Packet4i pandnot<Packet4i>(const Packet4i& a, const Packet4i& b) { return vec_and(a, vec_nor(b, b)); }
template<> EIGEN_STRONG_INLINE Packet4f pandnot<Packet4f>(const Packet4f& a, const Packet4f& b) { return vec_andc(a, b); }
template<> EIGEN_STRONG_INLINE Packet4i pandnot<Packet4i>(const Packet4i& a, const Packet4i& b) { return vec_andc(a, b); }
template<> EIGEN_STRONG_INLINE Packet4f pselect(const Packet4f& mask, const Packet4f& a, const Packet4f& b) {
return vec_sel(b, a, reinterpret_cast<Packet4ui>(mask));
@ -1260,15 +1260,15 @@ EIGEN_STRONG_INLINE Packet8bf F32ToBf16(Packet4f p4f){
Packet4bi is_max_exp = vec_cmpeq(exp, p4ui_max_exp);
Packet4bi is_zero_exp = vec_cmpeq(exp, reinterpret_cast<Packet4ui>(p4i_ZERO));
Packet4bi is_mant_not_zero = vec_cmpne(mantissa, reinterpret_cast<Packet4ui>(p4i_ZERO));
Packet4ui nan_selector = pand<Packet4ui>(
Packet4bi is_mant_zero = vec_cmpeq(mantissa, reinterpret_cast<Packet4ui>(p4i_ZERO));
Packet4ui nan_selector = pandnot<Packet4ui>(
reinterpret_cast<Packet4ui>(is_max_exp),
reinterpret_cast<Packet4ui>(is_mant_not_zero)
reinterpret_cast<Packet4ui>(is_mant_zero)
);
Packet4ui subnormal_selector = pand<Packet4ui>(
Packet4ui subnormal_selector = pandnot<Packet4ui>(
reinterpret_cast<Packet4ui>(is_zero_exp),
reinterpret_cast<Packet4ui>(is_mant_not_zero)
reinterpret_cast<Packet4ui>(is_mant_zero)
);
const _EIGEN_DECLARE_CONST_FAST_Packet4ui(nan, 0x7FC00000);