diff options
| author | Sanjay Patel <spatel@rotateright.com> | 2014-12-19 16:44:08 +0000 |
|---|---|---|
| committer | Sanjay Patel <spatel@rotateright.com> | 2014-12-19 16:44:08 +0000 |
| commit | ea3c8028871749a289d9d40a13962c0cf90d3c35 (patch) | |
| tree | 4054430c163a821c8006a9b74c1de2d8b9c5c7f6 /llvm/test/Transforms/InstCombine/fast-math.ll | |
| parent | 55ef1515021a9dea737bea38468950999945884a (diff) | |
| download | bcm5719-llvm-ea3c8028871749a289d9d40a13962c0cf90d3c35.tar.gz bcm5719-llvm-ea3c8028871749a289d9d40a13962c0cf90d3c35.zip | |
use -0.0 when creating an fneg instruction
Backends recognize (-0.0 - X) as the canonical form for fneg
and produce better code. Eg, ppc64 with 0.0:
lis r2, ha16(LCPI0_0)
lfs f0, lo16(LCPI0_0)(r2)
fsubs f1, f0, f1
blr
vs. -0.0:
fneg f1, f1
blr
Differential Revision: http://reviews.llvm.org/D6723
llvm-svn: 224583
Diffstat (limited to 'llvm/test/Transforms/InstCombine/fast-math.ll')
| -rw-r--r-- | llvm/test/Transforms/InstCombine/fast-math.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/test/Transforms/InstCombine/fast-math.ll b/llvm/test/Transforms/InstCombine/fast-math.ll index b0ec8952bb3..f39adba6560 100644 --- a/llvm/test/Transforms/InstCombine/fast-math.ll +++ b/llvm/test/Transforms/InstCombine/fast-math.ll @@ -93,7 +93,7 @@ define float @fold9(float %f1, float %f2) { ret float %t3 ; CHECK-LABEL: @fold9( -; CHECK: fsub fast float 0.000000e+00, %f2 +; CHECK: fsub fast float -0.000000e+00, %f2 } ; Let C3 = C1 + C2. (f1 + C1) + (f2 + C2) => (f1 + f2) + C3 instead of |
