AArch64: Add FULLY_PIPELINED_FMA to tune baseline
Add FULLY_PIPELINED_FMA to tune baseline - this is a generic feature that is already enabled for some cores, but benchmarking it shows it is faster on all modern cores (SPECFP improves ~0.17% on Neoverse V1 and 0.04% on Neoverse N1). gcc: * config/aarch64/aarch64-tuning-flags.def (AARCH64_EXTRA_TUNE_BASE): Add AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA. * config/aarch64/tuning_models/ampere1b.h: Remove redundant AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA. * config/aarch64/tuning_models/neoversev2.h: Likewise.
Showing
- gcc/config/aarch64/aarch64-tuning-flags.def 2 additions, 1 deletiongcc/config/aarch64/aarch64-tuning-flags.def
- gcc/config/aarch64/tuning_models/ampere1b.h 1 addition, 2 deletionsgcc/config/aarch64/tuning_models/ampere1b.h
- gcc/config/aarch64/tuning_models/neoversev2.h 1 addition, 2 deletionsgcc/config/aarch64/tuning_models/neoversev2.h
Loading
Please register or sign in to comment