diff options
| author | Sanjay Patel <spatel@rotateright.com> | 2018-05-21 23:54:19 +0000 |
|---|---|---|
| committer | Sanjay Patel <spatel@rotateright.com> | 2018-05-21 23:54:19 +0000 |
| commit | 17a870f07cf41372fe85c603e2338f605b3f196b (patch) | |
| tree | 5e14280f606e26f12354e2da6b810980eab6fc20 /llvm/test/CodeGen/NVPTX/implicit-def.ll | |
| parent | 5cb2e30c9805e175739cfd80a94f8022a8a84b64 (diff) | |
| download | bcm5719-llvm-17a870f07cf41372fe85c603e2338f605b3f196b.tar.gz bcm5719-llvm-17a870f07cf41372fe85c603e2338f605b3f196b.zip | |
[DAG] fold FP binops with undef operands to NaN
This is the FP sibling of D43141 with the corresponding IR change in rL327212.
We can't propagate undef here because if a variable operand is a NaN, these
binops must propagate NaN. Neither global nor node-level fast-math makes a
difference. If we have 'nnan', I think later folds can turn the NaN into undef.
The tests in X86/fp-undef.ll are meant to be the definitive verification for
these folds - everything reduces identically now.
The other test changes are collateral damage. They may need to be altered to
preserve their intent.
Differential Revision: https://reviews.llvm.org/D47026
llvm-svn: 332920
Diffstat (limited to 'llvm/test/CodeGen/NVPTX/implicit-def.ll')
| -rw-r--r-- | llvm/test/CodeGen/NVPTX/implicit-def.ll | 9 |
1 files changed, 0 insertions, 9 deletions
diff --git a/llvm/test/CodeGen/NVPTX/implicit-def.ll b/llvm/test/CodeGen/NVPTX/implicit-def.ll deleted file mode 100644 index 2d2c6e527f6..00000000000 --- a/llvm/test/CodeGen/NVPTX/implicit-def.ll +++ /dev/null @@ -1,9 +0,0 @@ -; RUN: llc < %s -O0 -march=nvptx -mcpu=sm_20 -asm-verbose=1 | FileCheck %s - -; CHECK: // implicit-def: %f[[F0:[0-9]+]] -; CHECK: add.rn.f32 %f{{[0-9]+}}, %f{{[0-9]+}}, %f[[F0]]; -define float @foo(float %a) { - %ret = fadd float %a, undef - ret float %ret -} - |
