procyberian/gcc
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

aarch64: Ignore inductions when costing vector code

Browse source 
In practice it seems to be better not to cost a vector induction.
The scalar code generally needs the same induction but doesn't
cost it, making an apples-for-apples comparison harder.  Most
inductions also have a low latency and their cost usually gets
hidden by other operations.

Like with the previous patches, this one only becomes active if
a CPU selects use_new_vector_costs.  It should therefore have
a very low impact on other CPUs.

gcc/
	* config/aarch64/aarch64.c (aarch64_detect_vector_stmt_subtype):
	Assume a zero cost for induction phis.
This commit is contained in:
Richard Sandiford 2021-03-26 16:08:37 +00:00
parent 99f94ae501
commit e4180ab2fe
1 changed files with 6 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
gcc/config/aarch64/aarch64.c
View file

@ -14541,6 +14541,12 @@ aarch64_detect_vector_stmt_subtype (vec_info *vinfo, vect_cost_for_stmt kind,
if (aarch64_sve_mode_p (TYPE_MODE (vectype)))
sve_costs = aarch64_tune_params.vec_costs->sve;
/* It's generally better to avoid costing inductions, since the induction
will usually be hidden by other operations. This is particularly true
for things like COND_REDUCTIONS. */
if (is_a<gphi *> (stmt_info->stmt))
return 0;
/* Detect cases in which vec_to_scalar is describing the extraction of a
vector element in preparation for a scalar store. The store itself is
costed separately. */