aboutsummaryrefslogtreecommitdiff
path: root/gcc
diff options
context:
space:
mode:
authorKewen Lin <linkw@linux.ibm.com>2021-09-21 22:25:54 -0500
committerKewen Lin <linkw@linux.ibm.com>2021-09-21 23:47:33 -0500
commit144c4984655e9c687dfdda6b1ac39afea8b2a037 (patch)
tree9ab6e1e1de670ebdb32745ef841a6a93374578f6 /gcc
parent72394d38d929c72cad6cf706a0e790b5934f2bf1 (diff)
downloadgcc-144c4984655e9c687dfdda6b1ac39afea8b2a037.zip
gcc-144c4984655e9c687dfdda6b1ac39afea8b2a037.tar.gz
gcc-144c4984655e9c687dfdda6b1ac39afea8b2a037.tar.bz2
rs6000: Parameterize some const values for density test
This patch follows the discussion here[1], where Segher suggested parameterizing those exact magic constants for density heuristics, to make it easier to tweak if need. The change here should be "No Functional Change". But I verified it with SPEC2017 at option sets O2-vect and Ofast-unroll on Power8, the result is neutral as expected. [1]https://gcc.gnu.org/pipermail/gcc-patches/2021-September/579121.html gcc/ChangeLog: * config/rs6000/rs6000.opt (rs6000-density-pct-threshold, rs6000-density-size-threshold, rs6000-density-penalty, rs6000-density-load-pct-threshold, rs6000-density-load-num-threshold): New parameter. * config/rs6000/rs6000.c (rs6000_density_test): Adjust with corresponding parameters.
Diffstat (limited to 'gcc')
-rw-r--r--gcc/config/rs6000/rs6000.c22
-rw-r--r--gcc/config/rs6000/rs6000.opt38
2 files changed, 45 insertions, 15 deletions
diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index 060f51a..ad86072 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6000.c
@@ -5289,9 +5289,6 @@ struct rs6000_cost_data
static void
rs6000_density_test (rs6000_cost_data *data)
{
- const int DENSITY_PCT_THRESHOLD = 85;
- const int DENSITY_SIZE_THRESHOLD = 70;
- const int DENSITY_PENALTY = 10;
struct loop *loop = data->loop_info;
basic_block *bbs = get_loop_body (loop);
int nbbs = loop->num_nodes;
@@ -5327,26 +5324,21 @@ rs6000_density_test (rs6000_cost_data *data)
free (bbs);
density_pct = (vec_cost * 100) / (vec_cost + not_vec_cost);
- if (density_pct > DENSITY_PCT_THRESHOLD
- && vec_cost + not_vec_cost > DENSITY_SIZE_THRESHOLD)
+ if (density_pct > rs6000_density_pct_threshold
+ && vec_cost + not_vec_cost > rs6000_density_size_threshold)
{
- data->cost[vect_body] = vec_cost * (100 + DENSITY_PENALTY) / 100;
+ data->cost[vect_body] = vec_cost * (100 + rs6000_density_penalty) / 100;
if (dump_enabled_p ())
dump_printf_loc (MSG_NOTE, vect_location,
"density %d%%, cost %d exceeds threshold, penalizing "
- "loop body cost by %d%%\n", density_pct,
- vec_cost + not_vec_cost, DENSITY_PENALTY);
+ "loop body cost by %u%%\n", density_pct,
+ vec_cost + not_vec_cost, rs6000_density_penalty);
}
/* Check whether we need to penalize the body cost to account
for excess strided or elementwise loads. */
if (data->extra_ctor_cost > 0)
{
- /* Threshold for load stmts percentage in all vectorized stmts. */
- const int DENSITY_LOAD_PCT_THRESHOLD = 45;
- /* Threshold for total number of load stmts. */
- const int DENSITY_LOAD_NUM_THRESHOLD = 20;
-
gcc_assert (data->nloads <= data->nstmts);
unsigned int load_pct = (data->nloads * 100) / data->nstmts;
@@ -5360,8 +5352,8 @@ rs6000_density_test (rs6000_cost_data *data)
the loads.
One typical case is the innermost loop of the hotspot of SPEC2017
503.bwaves_r without loop interchange. */
- if (data->nloads > DENSITY_LOAD_NUM_THRESHOLD
- && load_pct > DENSITY_LOAD_PCT_THRESHOLD)
+ if (data->nloads > (unsigned int) rs6000_density_load_num_threshold
+ && load_pct > (unsigned int) rs6000_density_load_pct_threshold)
{
data->cost[vect_body] += data->extra_ctor_cost;
if (dump_enabled_p ())
diff --git a/gcc/config/rs6000/rs6000.opt b/gcc/config/rs6000/rs6000.opt
index c1cb9ab..9d7878f 100644
--- a/gcc/config/rs6000/rs6000.opt
+++ b/gcc/config/rs6000/rs6000.opt
@@ -639,3 +639,41 @@ Enable instructions that guard against return-oriented programming attacks.
mprivileged
Target Var(rs6000_privileged) Init(0)
Generate code that will run in privileged state.
+
+-param=rs6000-density-pct-threshold=
+Target Undocumented Joined UInteger Var(rs6000_density_pct_threshold) Init(85) IntegerRange(0, 100) Param
+When costing for loop vectorization, we probably need to penalize the loop body
+cost if the existing cost model may not adequately reflect delays from
+unavailable vector resources. We collect the cost for vectorized statements
+and non-vectorized statements separately, check the proportion of vec_cost to
+total cost of vec_cost and non vec_cost, and penalize only if the proportion
+exceeds the threshold specified by this parameter. The default value is 85.
+
+-param=rs6000-density-size-threshold=
+Target Undocumented Joined UInteger Var(rs6000_density_size_threshold) Init(70) IntegerRange(0, 1000) Param
+Like parameter rs6000-density-pct-threshold, we also check the total sum of
+vec_cost and non vec_cost, and penalize only if the sum exceeds the threshold
+specified by this parameter. The default value is 70.
+
+-param=rs6000-density-penalty=
+Target Undocumented Joined UInteger Var(rs6000_density_penalty) Init(10) IntegerRange(0, 1000) Param
+When both heuristics with rs6000-density-pct-threshold and
+rs6000-density-size-threshold are satisfied, we decide to penalize the loop
+body cost by the value which is specified by this parameter. The default
+value is 10.
+
+-param=rs6000-density-load-pct-threshold=
+Target Undocumented Joined UInteger Var(rs6000_density_load_pct_threshold) Init(45) IntegerRange(0, 100) Param
+When costing for loop vectorization, we probably need to penalize the loop body
+cost by accounting for excess strided or elementwise loads. We collect the
+numbers for general statements and load statements according to the information
+for statements to be vectorized, check the proportion of load statements, and
+penalize only if the proportion exceeds the threshold specified by this
+parameter. The default value is 45.
+
+-param=rs6000-density-load-num-threshold=
+Target Undocumented Joined UInteger Var(rs6000_density_load_num_threshold) Init(20) IntegerRange(0, 1000) Param
+Like parameter rs6000-density-load-pct-threshold, we also check if the total
+number of load statements exceeds the threshold specified by this parameter,
+and penalize only if it's satisfied. The default value is 20.
+