Add support for vector permute cost since various permutes can expand into a…
Add support for vector permute cost since various permutes can expand into a complex sequence of instructions. Add support for vector permute cost since various permutes can expand into a complex sequence of instructions. This fixes major performance regressions due to recent changes in SLP vectorizer (which now vectorizes more aggressively and emits many complex permutes). Set the cost to > 1 for all microarchitectures so that the number of permutes is usually zero and regressions disappear. 2016-01-28 Wilco Dijkstra <wdijkstr@arm.com> * config/aarch64/aarch64.c (generic_vector_cost): Set vec_permute_cost. (cortexa57_vector_cost): Likewise. (exynosm1_vector_cost): Likewise. (xgene1_vector_cost): Likewise. (aarch64_builtin_vectorization_cost): Use vec_permute_cost. * config/aarch64/aarch64-protos.h (cpu_vector_cost): Add vec_permute_cost entry. From-SVN: r232922
Showing
Please
register
or
sign in
to comment