From e9c4fbe9c1eff61ffe2aa31c67cef2ffd346f674 Mon Sep 17 00:00:00 2001 From: Bernd Schmidt Date: Mon, 23 Jan 2017 16:17:33 +0000 Subject: [PATCH] re PR rtl-optimization/78634 (30% performance drop after r242832.) PR rtl-optimization/78634 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function. (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define. * ifcvt.c (noce_try_cmove): Add missing cost check. testsuite/ PR rtl-optimization/78634 * gcc.target/i386/funcspec-11.c: Also pass -mtune=i686. From-SVN: r244816 --- gcc/ChangeLog | 7 ++++++ gcc/config/i386/i386.c | 24 +++++++++++++++++++++ gcc/ifcvt.c | 2 +- gcc/testsuite/ChangeLog | 5 +++++ gcc/testsuite/gcc.target/i386/funcspec-11.c | 2 +- 5 files changed, 38 insertions(+), 2 deletions(-) diff --git a/gcc/ChangeLog b/gcc/ChangeLog index bd716492ec1..35694c29b99 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,10 @@ +2017-01-23 Bernd Schmidt + + PR rtl-optimization/78634 + * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function. + (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define. + * ifcvt.c (noce_try_cmove): Add missing cost check. + 2017-01-23 Bin Cheng PR tree-optimization/70754 diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c index 65b32e6da56..15498a954ea 100644 --- a/gcc/config/i386/i386.c +++ b/gcc/config/i386/i386.c @@ -50486,6 +50486,28 @@ ix86_spill_class (reg_class_t rclass, machine_mode mode) return NO_REGS; } +/* Implement TARGET_MAX_NOCE_IFCVT_SEQ_COST. Like the default implementation, + but returns a lower bound. */ + +static unsigned int +ix86_max_noce_ifcvt_seq_cost (edge e) +{ + bool predictable_p = predictable_edge_p (e); + + enum compiler_param param + = (predictable_p + ? PARAM_MAX_RTL_IF_CONVERSION_PREDICTABLE_COST + : PARAM_MAX_RTL_IF_CONVERSION_UNPREDICTABLE_COST); + + /* If we have a parameter set, use that, otherwise take a guess using + BRANCH_COST. */ + if (global_options_set.x_param_values[param]) + return PARAM_VALUE (param); + else + return BRANCH_COST (true, predictable_p) * COSTS_N_INSNS (2); +} + + /* Implement targetm.vectorize.init_cost. */ static void * @@ -52035,6 +52057,8 @@ ix86_run_selftests (void) #undef TARGET_EXPAND_DIVMOD_LIBFUNC #define TARGET_EXPAND_DIVMOD_LIBFUNC ix86_expand_divmod_libfunc +#undef TARGET_MAX_NOCE_IFCVT_SEQ_COST +#define TARGET_MAX_NOCE_IFCVT_SEQ_COST ix86_max_noce_ifcvt_seq_cost #if CHECKING_P #undef TARGET_RUN_TARGET_SELFTESTS #define TARGET_RUN_TARGET_SELFTESTS selftest::ix86_run_selftests diff --git a/gcc/ifcvt.c b/gcc/ifcvt.c index 6d306392f60..e2d6c6fcc51 100644 --- a/gcc/ifcvt.c +++ b/gcc/ifcvt.c @@ -1826,7 +1826,7 @@ noce_try_cmove (struct noce_if_info *if_info) noce_emit_move_insn (if_info->x, target); seq = end_ifcvt_sequence (if_info); - if (!seq) + if (!seq || !noce_conversion_profitable_p (seq, if_info)) return FALSE; emit_insn_before_setloc (seq, if_info->jump, diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog index de1f24b5e03..5bdeda87223 100644 --- a/gcc/testsuite/ChangeLog +++ b/gcc/testsuite/ChangeLog @@ -1,3 +1,8 @@ +2017-01-23 Bernd Schmidt + + PR rtl-optimization/78634 + * gcc.target/i386/funcspec-11.c: Also pass -mtune=i686. + 2017-01-23 Bin Cheng PR tree-optimization/70754 diff --git a/gcc/testsuite/gcc.target/i386/funcspec-11.c b/gcc/testsuite/gcc.target/i386/funcspec-11.c index 7c39f4cd2a1..fc79ac50742 100644 --- a/gcc/testsuite/gcc.target/i386/funcspec-11.c +++ b/gcc/testsuite/gcc.target/i386/funcspec-11.c @@ -1,7 +1,7 @@ /* PR target/36936 */ /* { dg-do compile } */ /* { dg-require-effective-target ia32 } */ -/* { dg-options "-O2 -march=i386" } */ +/* { dg-options "-O2 -march=i386 -mtune=i686" } */ /* { dg-final { scan-assembler "cmov" } } */ extern int foo (int) __attribute__((__target__("arch=i686"))); -- 2.30.2