Perform manual loop fission in bayestar_sky_map_toa_phoa_snr_pixel
Rearrange loops and rewrite one key loop with explicit SIMD intrinsics. This is seen to provide a speedup by a factor of 2 on Intel Skylake machines.
Edited by Leo P. Singer
Rearrange loops and rewrite one key loop with explicit SIMD intrinsics. This is seen to provide a speedup by a factor of 2 on Intel Skylake machines.