mirror of
https://gitlab.com/libeigen/eigen.git
synced 2026-04-10 11:34:33 +08:00
Use the correct truncating intrinsic for double->int casting.
This commit is contained in:
@@ -101,8 +101,8 @@ template<> EIGEN_STRONG_INLINE Packet4f pcast<Packet2d, Packet4f>(const Packet2d
|
||||
}
|
||||
|
||||
template<> EIGEN_STRONG_INLINE Packet4i pcast<Packet2d, Packet4i>(const Packet2d& a, const Packet2d& b) {
|
||||
return _mm_castps_si128(_mm_shuffle_ps(_mm_castsi128_ps(_mm_cvtpd_epi32(a)),
|
||||
_mm_castsi128_ps(_mm_cvtpd_epi32(b)),
|
||||
return _mm_castps_si128(_mm_shuffle_ps(_mm_castsi128_ps(_mm_cvttpd_epi32(a)),
|
||||
_mm_castsi128_ps(_mm_cvttpd_epi32(b)),
|
||||
(1 << 2) | (1 << 6)));
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user