This little series provides some cleanups for opt_algebraic. The most important of which is adding descriptions to a bunch of the asserts in the bit-size checker. I've only gotten through about half of it so there's still more work to do but it should make the error messages printed out much less opaque to the average reader. I can't count the number of times in the last year or so that someone has come to me asking why their obviously correct optimization fails validation. Hopefully, this will make working on nir_opt_algebraic even easier.
Cc: Connor Abbott <cwabbo...@gmail.com> Cc: Ian Romanick <i...@freedesktop.org> Jason Ekstrand (6): nir/algebraic: Use bool internally instead of bool32 nir/algebraic: Generalize an optimization nir/algebraic: Make internal classes str-able nir/algebraic: A bit of validation refactoring nir/algebraic: Losten a restriction on variables nir/algebraic: Provide descriptive asserts for bit size checks src/compiler/nir/nir_algebraic.py | 107 ++++++++++++++++++-------- src/compiler/nir/nir_opt_algebraic.py | 2 +- src/compiler/nir/nir_search.c | 6 +- 3 files changed, 78 insertions(+), 37 deletions(-) -- 2.19.1 _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
