[AMDGPU] Run always inliner early in opt

We can mark functions to always inline early in the opt. Since we do not have
call support this early inlining creates opportunities for inter-procedural
optimizations which would not occur otherwise.

Differential Revision: https://reviews.llvm.org/D31016

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297958 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
index 5e633a3ca376ff0ce40023a88bbb067ae499ea99..f5973c174985b29c67b30f1e4ff1feaca4daa294 100644 (file)
--- a/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
+++ b/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
@@ -244,6 +244,7 @@ void AMDGPUTargetMachine::adjustPassManager(PassManagerBuilder &Builder) {
         }));
         PM.add(createGlobalDCEPass());
       }
+      PM.add(createAMDGPUAlwaysInlinePass());
   });
 }
 

diff --git a/test/CodeGen/AMDGPU/early-inline.ll b/test/CodeGen/AMDGPU/early-inline.ll
new file mode 100644 (file)
index 0000000..dd526da
--- /dev/null
+++ b/test/CodeGen/AMDGPU/early-inline.ll
@@ -0,0 +1,21 @@
+; RUN: opt -mtriple=amdgcn-- -O1 -S -inline-threshold=1 %s | FileCheck %s
+
+define i32 @callee(i32 %x) {
+entry:
+  %mul1 = mul i32 %x, %x
+  %mul2 = mul i32 %mul1, %x
+  %mul3 = mul i32 %mul1, %mul2
+  %mul4 = mul i32 %mul3, %mul2
+  %mul5 = mul i32 %mul4, %mul3
+  ret i32 %mul5
+}
+
+; CHECK-LABEL: @caller
+; CHECK: mul i32
+; CHECK-NOT: call i32
+
+define i32 @caller(i32 %x) {
+entry:
+  %res = call i32 @callee(i32 %x)
+  ret i32 %res
+}