improve x86 support (fixes #342) #540

Job	Run time
Test installation using system supplied libs	33s
Python 3.7	8m 20s
Python 3.7	10m 51s
Python 3.7	11m 39s
Python 3.8	7m 58s
Python 3.8	8m 28s
Python 3.8	12m 33s
Python 3.9	7m 9s
Python 3.9	8m 51s
Python 3.9	11m 33s
Python 3.10	5m 53s
Python 3.10	10m 6s
Python 3.10	17m 22s
Python 3.11	6m 38s
Python 3.11	10m 46s
Python 3.11	14m 15s
Test install with generated cython files	4m 18s
Test sse2 code on Linux (gcc)	11m 24s
Test installation using newest version of Taskflow	29s
	2h 49m 6s

Provide feedback