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

Fixes #25

Vectorization improvements

See merge request pycodegen/pystencils!228

Fix installation from git

Closes #14

See merge request pycodegen/pystencils!227

Fix field size

Closes #32, #31, #3, and #7

See merge request !226

Fix benchmark generation

See merge request !223

ARM Neon: Fix makeVec and add Philox

See merge request !222

Ship c-file with pypi

See merge request !221

Improve ARM64 support

See merge request !220

Make RNG vectorization more robust

See merge request !219

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

The vectorized RNG can switch between scalar and vector types due to loop cutting.

Update setup.py

See merge request pycodegen/pystencils!218

Add kerncraft templates

See merge request !217

some fixes for lbmpy vectorization

See merge request pycodegen/pystencils!216

fix aligned_alloc on windows

See merge request !215

default to C++17 on macOS and Windows

Closes #24

See merge request !214

vectorization: improve treatment of unary minus

See merge request !212

Fix clear cache on startup

See merge request !213

Due two weird recursive dependencies `clear_cache` will not be
available at this point. In-lining this function will make
the config key `cache.clear_cache_on_start: true` work again.