[Patch 02/02] Introduce prefetch-dynamic-strides option
authorLuis Machado <luis.machado@linaro.org>
Wed, 23 May 2018 16:23:11 +0000 (16:23 +0000)
committerLuis Machado <luisgpm@gcc.gnu.org>
Wed, 23 May 2018 16:23:11 +0000 (16:23 +0000)
commitd2ff35c0913539a31782fbc05bb6d4c2ce74f2ce
tree36f27cdd1a5fcc65a32d45f16e798bdd393b5cb3
parent59100dfc42bbe92caff61bca1560da4a30f99906
[Patch 02/02] Introduce prefetch-dynamic-strides option

The following patch adds an option to control software prefetching of memory
references with non-constant/unknown strides.

Currently we prefetch these references if the pass thinks there is benefit to
doing so. But, since this is all based on heuristics, it's not always the case
that we end up with better performance.

For Falkor there is also the problem of conflicts with the hardware prefetcher,
so we need to be more conservative in terms of what we issue software prefetch
hints for.

This also aligns GCC with what LLVM does for Falkor.

Similarly to the previous patch, the defaults guarantee no change in behavior
for other targets and architectures.

gcc/ChangeLog:

2018-05-23  Luis Machado  <luis.machado@linaro.org>

* config/aarch64/aarch64-protos.h (cpu_prefetch_tune)
<prefetch_dynamic_strides>: New const bool field.
* config/aarch64/aarch64.c (generic_prefetch_tune): Update to include
prefetch_dynamic_strides.
(exynosm1_prefetch_tune): Likewise.
(thunderxt88_prefetch_tune): Likewise.
(thunderx_prefetch_tune): Likewise.
(thunderx2t99_prefetch_tune): Likewise.
(qdf24xx_prefetch_tune): Likewise. Set prefetch_dynamic_strides to
false.
(aarch64_override_options_internal): Update to set
PARAM_PREFETCH_DYNAMIC_STRIDES.
* doc/invoke.texi (prefetch-dynamic-strides): Document new option.
* params.def (PARAM_PREFETCH_DYNAMIC_STRIDES): New.
* params.h (PARAM_PREFETCH_DYNAMIC_STRIDES): Define.
* tree-ssa-loop-prefetch.c (should_issue_prefetch_p): Account for
prefetch-dynamic-strides setting.

From-SVN: r260618
gcc/ChangeLog
gcc/config/aarch64/aarch64-protos.h
gcc/config/aarch64/aarch64.c
gcc/doc/invoke.texi
gcc/params.def
gcc/params.h
gcc/tree-ssa-loop-prefetch.c