diff options
| author | Zvi Rackover <zvi.rackover@intel.com> | 2016-09-26 06:42:07 +0000 |
|---|---|---|
| committer | Zvi Rackover <zvi.rackover@intel.com> | 2016-09-26 06:42:07 +0000 |
| commit | 839d15a194a658dfe407e1681514231455138360 (patch) | |
| tree | 554c7066bdf9216a5962edd93e21927bf4d9705c /llvm/test/CodeGen/X86/pr29022.ll | |
| parent | 9d68b64c00e43a5b5430c5cdbd78f4ae85aab0b4 (diff) | |
| download | bcm5719-llvm-839d15a194a658dfe407e1681514231455138360.tar.gz bcm5719-llvm-839d15a194a658dfe407e1681514231455138360.zip | |
[X86] Optimization for replacing LEA with MOV at frame index elimination time
Summary:
Replace a LEA instruction of the form 'lea (%esp), %ebx' --> 'mov %esp, %ebx'
MOV is preferable over LEA because usually there are more issue-slots available to execute MOVs than LEAs. Latest processors also support zero-latency MOVs.
Fixes pr29022.
Reviewers: hfinkel, delena, igorb, myatsina, mkuper
Differential Revision: https://reviews.llvm.org/D24705
llvm-svn: 282385
Diffstat (limited to 'llvm/test/CodeGen/X86/pr29022.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/pr29022.ll | 15 |
1 files changed, 15 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/X86/pr29022.ll b/llvm/test/CodeGen/X86/pr29022.ll new file mode 100644 index 00000000000..85bfc7f1118 --- /dev/null +++ b/llvm/test/CodeGen/X86/pr29022.ll @@ -0,0 +1,15 @@ +; RUN: llc < %s -mcpu=skx -mtriple x86_64-unknown-linux-gnu -verify-machineinstrs | FileCheck %s +; RUN: llc < %s -mcpu=skx -mtriple=x86_64-linux-gnux32 -verify-machineinstrs | FileCheck %s --check-prefix=X32 + +define i32 @A() { +; CHECK: movq %rsp, %rdi +; CHECK-NEXT: call + +; X32: movl %esp, %edi +; X32-NEXT: call + %alloc = alloca i32, align 8 + %call = call i32 @foo(i32* %alloc) + ret i32 %call +} + +declare i32 @foo(i32*) |
