simdpp::unzip16_hi, unzip8_hi, unzip4_hi, unzip2_hi
From libsimdpp-docs
< shuffle
template<unsigned N, class V1, class V2>
Ret<N, _DETAIL_> unzip16_hi( const any_vec8<N,V1>& a, |
||
template<unsigned N, class V1, class V2>
Ret<N, _DETAIL_> unzip8_hi( const any_vec16<N,V1>& a, |
||
template<unsigned N, class V1, class V2>
Ret<N, _DETAIL_> unzip4_hi( const any_vec32<N,V1>& a, |
||
template<unsigned N, class V1, class V2>
Ret<N, _DETAIL_> unzip2_hi( const any_vec64<N,V1>& a, |
||
De-interleaves odd elements from two source vectors. Each 128-bit lane within vectors is processed separately.
The return vector type is governed by the expression promotion rules.
Parameters[edit]
a, b | - | source vectors |
Return value[edit]
Equivalent operation[edit]
For the uint32x4 case:
r = [ a1 a3 b1 b3 ]
For the uint32x8 case:
r = [ a1 a3 b1 b3 a5 a7 b5 b7 ]
See also[edit]
This section is incomplete |