- fix the aligned version
- make sure the test case is incommensurate with the vector width
- implement a fallback for instruction sets that don't support it natively

otherwise we can delete stuff from the next outer loop, which in 2D would be the one OpenMP-parallelized

Fixes #25

Michael Kuron authored 4 years ago and Markus Holzer committed 4 years ago

and clean up some old xfails

Stephan Seitz authored 4 years ago and Markus Holzer committed 4 years ago

We have to try from newest to oldest import to avoid deprecation
warnings.