Add matrixMultiplyAVX2b

This version is based on matrixMultiplyAVX2 with some changes:

• Remove the multiplication with inv
• Use _mm256_addsub_ps
• Replace the overkill _mm256_permutevar8x32_ps with cheaper (and cleaner) _mm256_permute_ps
• Reshuffle b_1 and b_2 to get b_3 and b_4

requested review from @mancini

assigned to @veenboer

added 9 commits

• 3d579fa9...a418699e - 3 commits from branch main
• 8c061688 - Add matrixMultiplyAVX2b
• 8a3c26ec - Remove multiplication with inv
• b25351c9 - Replace add and mul with fmadd
• 78b92033 - Simplify using _mm256_permute_ps
• ac90b5bf - Remove helper variables
• e1a55db4 - Use permute_ps to swap real and imag

Compare with previous version

approved this merge request

enabled an automatic merge when the pipeline for e1a55db4 succeeds

mentioned in commit 815ae398

merged