A high-performance general-purpose MRF MAP solver, heavily exploiting SIMD instructions.
BSD 3-Clause "New" or "Revised" License
102
stars
51
forks
source link
Removed AVX instruction `_mm_maskstore_ps` from SSE wrapper #8
Closed
magcks closed 6 years ago
I replaced the
_mm_maskstore_ps
intrinsic from all SSE masked_store functions with a combination of two_mm_maskmove_si64
.The code can be tested with the following 'tests':