diff options
| author | Eric Christopher <echristo@gmail.com> | 2019-02-20 04:42:07 +0000 |
|---|---|---|
| committer | Eric Christopher <echristo@gmail.com> | 2019-02-20 04:42:07 +0000 |
| commit | 2534592b9fe436004fa7fd7da73cbb3a00902b9c (patch) | |
| tree | 18fa4ff46b46043b2859cf03446f62c35a3792eb /llvm/lib/Target/X86/X86TargetTransformInfo.h | |
| parent | d990c2a9e23f1b8d6d7ee17c2d1f46202d70b88e (diff) | |
| download | bcm5719-llvm-2534592b9fe436004fa7fd7da73cbb3a00902b9c.tar.gz bcm5719-llvm-2534592b9fe436004fa7fd7da73cbb3a00902b9c.zip | |
Temporarily Revert "[X86][SLP] Enable SLP vectorization for 128-bit horizontal X86 instructions (add, sub)"
As this has broken the lto bootstrap build for 3 days and is
showing a significant regression on the Dither_benchmark results (from
the LLVM benchmark suite) -- specifically, on the
BENCHMARK_FLOYD_DITHER_128, BENCHMARK_FLOYD_DITHER_256, and
BENCHMARK_FLOYD_DITHER_512; the others are unchanged. These have
regressed by about 28% on Skylake, 34% on Haswell, and over 40% on
Sandybridge.
This reverts commit r353923.
llvm-svn: 354434
Diffstat (limited to 'llvm/lib/Target/X86/X86TargetTransformInfo.h')
| -rw-r--r-- | llvm/lib/Target/X86/X86TargetTransformInfo.h | 1 |
1 files changed, 0 insertions, 1 deletions
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.h b/llvm/lib/Target/X86/X86TargetTransformInfo.h index 36a941da66d..8bfcd07be04 100644 --- a/llvm/lib/Target/X86/X86TargetTransformInfo.h +++ b/llvm/lib/Target/X86/X86TargetTransformInfo.h @@ -113,7 +113,6 @@ public: unsigned getNumberOfRegisters(bool Vector); unsigned getRegisterBitWidth(bool Vector) const; - unsigned getMinVectorRegisterBitWidth() const; unsigned getLoadStoreVecRegBitWidth(unsigned AS) const; unsigned getMaxInterleaveFactor(unsigned VF); int getArithmeticInstrCost( |
