Better overload #366

Job	Run time
macosx (clang, x86_osx, -mavx)	21m 59s
x86-clang (clang, x86, -msse2)	14m 29s
macosx (gcc, x86_osx, -mavx)	22m 0s
no_pch (clang, x86, -msse2)	21m 52s
x86-clang (clang, x86, -msse4.2)	14m 19s
avx512 (clang, x86, -march=skylake-avx512)	20m 50s
x86-clang (clang, x86, -mavx)	16m 52s
avx512 (clang, x86, -DEVE_NO_SIMD)	15m 56s
x86-clang (clang, x86, -mavx2)	16m 32s
avx512 (clang, x86_asan, -march=skylake-avx512)	1m 14s
avx512 (gcc, x86, -march=skylake-avx512)	1s
avx512 (gcc, x86, -DEVE_NO_SIMD)	6m 1s
other-arch (gcc, aarch64, -Wno-psabi)	4m 59s
other-arch (gcc, arm, -Wno-psabi)	4m 59s
other-arch (gcc, ppc64, -Wno-psabi)	4m 59s
x86-gcc	1s
x86-custom	1s
arm-sve	1s
	3h 7m 5s

Provide feedback