SSE should always work! AVX is enabled automatically if it is available. To check availability have a look at this list of defines
To check if the defines work as expected even though the own computer does not support them (-march=native), one can add the following compile flags: -mavx2 -mavx512f
[x] Remove SSE related options from CMake scripts
[x] Add #ifdefto source files if needed
[x] Put the fused multiply add functions in a '#ifndef`#else block (check version number for that).
[x] Implement fused multiply add for SSE version that do not support this operation. It is just a combination of _mm_mul and _mm_add
SSE should always work! AVX is enabled automatically if it is available. To check availability have a look at this list of defines
To check if the defines work as expected even though the own computer does not support them (
-march=native
), one can add the following compile flags:-mavx2 -mavx512f
#ifdef
to source files if needed`#else
block (check version number for that)._mm_mul
and_mm_add