Disable the tune for Zhaoxin/CLX/SKX since it could hurt performance
for the inner loop.
According to last test, align_loop helps performance for SPEC2017 on EMR and
Znver4.
So I'll still keep the tune for generic part.
Bootstrapped and regtested on x86_64-pc-linux-gnu{-m32,}.
Any comment?
gcc/ChangeLog:
PR target/117438
* config/i386/i386-features.cc (pass_align_tight_loops::gate):
Add TARGET_ALIGN_TIGHT_LOOPS to the predicate.
* config/i386/i386.h (TARGET_ALIGN_TIGHT_LOOPS): New macro.
* config/i386/x86-tune.def (X86_TUNE_ALIGN_TIGHT_LOOPS): New
tune.
---
gcc/config/i386/i386-features.cc | 3 ++-
gcc/config/i386/i386.h | 2 ++
gcc/config/i386/x86-tune.def | 8 +++++++-
3 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/gcc/config/i386/i386-features.cc b/gcc/config/i386/i386-features.cc
index e2e85212a4f..70bda4bc021 100644
--- a/gcc/config/i386/i386-features.cc
+++ b/gcc/config/i386/i386-features.cc
@@ -3620,7 +3620,8 @@ public:
/* opt_pass methods: */
bool gate (function *) final override
{
- return optimize && optimize_function_for_speed_p (cfun);
+ return TARGET_ALIGN_TIGHT_LOOPS
+ && optimize && optimize_function_for_speed_p (cfun);
}
unsigned int execute (function *) final override
diff --git a/gcc/config/i386/i386.h b/gcc/config/i386/i386.h
index 51934400951..2c6dbf6dfdc 100644
--- a/gcc/config/i386/i386.h
+++ b/gcc/config/i386/i386.h
@@ -466,6 +466,8 @@ extern unsigned char ix86_tune_features[X86_TUNE_LAST];
#define TARGET_USE_RCR ix86_tune_features[X86_TUNE_USE_RCR]
#define TARGET_SSE_MOVCC_USE_BLENDV \
ix86_tune_features[X86_TUNE_SSE_MOVCC_USE_BLENDV]
+#define TARGET_ALIGN_TIGHT_LOOPS \
+ ix86_tune_features[X86_TUNE_ALIGN_TIGHT_LOOPS]
/* Feature tests against the various architecture variations. */
enum ix86_arch_indices {
diff --git a/gcc/config/i386/x86-tune.def b/gcc/config/i386/x86-tune.def
index 6ebb2fd3414..8afa0cd9823 100644
--- a/gcc/config/i386/x86-tune.def
+++ b/gcc/config/i386/x86-tune.def
@@ -214,7 +214,7 @@ DEF_TUNE (X86_TUNE_SINGLE_POP, "single_pop", m_386 | m_486
| m_PENT
DEF_TUNE (X86_TUNE_DOUBLE_POP, "double_pop", m_PENT | m_LAKEMONT)
/*****************************************************************************/
-/* Branch predictor tuning */
+/* Branch predictor and The Front-end tuning
*/
/*****************************************************************************/
/* X86_TUNE_PAD_SHORT_FUNCTION: Make every function to be at least 4
@@ -235,6 +235,12 @@ DEF_TUNE (X86_TUNE_FOUR_JUMP_LIMIT, "four_jump_limit",
m_PPRO | m_P4_NOCONA | m_BONNELL | m_SILVERMONT | m_GOLDMONT
| m_GOLDMONT_PLUS | m_INTEL | m_ATHLON_K8 | m_AMDFAM10)
+/* X86_TUNE_ALIGN_TIGHT_LOOP: For tight loop whose size is
+ smaller than prefetch_block, align it to ceil_log2 (loop_size). The tune
+ overwrites -falign-loops=N. */
+DEF_TUNE (X86_TUNE_ALIGN_TIGHT_LOOPS, "align_tight_loops",
+ ~(m_ZHAOXIN | m_CASCADELAKE | m_SKYLAKE_AVX512))
+
/*****************************************************************************/
/* Integer instruction selection tuning */
/*****************************************************************************/
--
2.34.1
