diff options
| author | Nadav Rotem <nrotem@apple.com> | 2013-04-12 00:48:32 +0000 |
|---|---|---|
| committer | Nadav Rotem <nrotem@apple.com> | 2013-04-12 00:48:32 +0000 |
| commit | c3b0f50ac25b2277ae8712abc03485923c09716c (patch) | |
| tree | 5299d0a1d862c399dfe4b1c796d944dca67483af /llvm/test/CodeGen/X86/code_placement_align_all.ll | |
| parent | 73b75e01bfa15ab4d308d05861e39a4f93cb15e9 (diff) | |
| download | bcm5719-llvm-c3b0f50ac25b2277ae8712abc03485923c09716c.tar.gz bcm5719-llvm-c3b0f50ac25b2277ae8712abc03485923c09716c.zip | |
Add a flag to align all basic blocks in the function.
When debugging performance regressions we often ask ourselves if the regression
that we see is due to poor isel/sched/ra or due to some micro-architetural
problem. When comparing two code sequences one good way to rule out front-end
bottlenecks (and other the issues) is to force code alignment. This pass adds
a flag that forces the alignment of all of the basic blocks in the program.
llvm-svn: 179353
Diffstat (limited to 'llvm/test/CodeGen/X86/code_placement_align_all.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/code_placement_align_all.ll | 22 |
1 files changed, 22 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/X86/code_placement_align_all.ll b/llvm/test/CodeGen/X86/code_placement_align_all.ll new file mode 100644 index 00000000000..2a36a3a3af5 --- /dev/null +++ b/llvm/test/CodeGen/X86/code_placement_align_all.ll @@ -0,0 +1,22 @@ +; RUN: llc -march=x86 -align-all-blocks=16 < %s | FileCheck %s + +;CHECK: foo +;CHECK: .align 16, 0x90 +;CHECK: .align 16, 0x90 +;CHECK: .align 16, 0x90 +;CHECK: ret +define i32 @foo(i32 %t, i32 %l) nounwind readnone ssp uwtable { + %1 = icmp eq i32 %t, 0 + br i1 %1, label %4, label %2 + +; <label>:2 ; preds = %0 + %3 = add nsw i32 %t, 2 + ret i32 %3 + +; <label>:4 ; preds = %0 + %5 = icmp eq i32 %l, 0 + %. = select i1 %5, i32 0, i32 5 + ret i32 %. +} + + |
