[llvm] 7c724a8 - [AMDGPU] Do not check max-bb for a single block callee

[Stanislav Mekhanoshin via llvm-commits]

Author: Stanislav Mekhanoshin
Date: 2021-03-01T19:48:50-08:00
New Revision: 7c724a896f93c97fe75db6f37b0995c9b35e0b82

URL: https://github.com/llvm/llvm-project/commit/7c724a896f93c97fe75db6f37b0995c9b35e0b82
DIFF: https://github.com/llvm/llvm-project/commit/7c724a896f93c97fe75db6f37b0995c9b35e0b82.diff

LOG: [AMDGPU] Do not check max-bb for a single block callee

-amdgpu-inline-max-bb option could lead to a suboptimal
codegen preventing inlining of really simple functions
including pure wrapper calls. Relax the cutoff by allowing
to call a function with a single block on the grounds
that it will not increase total number of blocks after
inlining.

Differential Revision: https://reviews.llvm.org/D97744

Added: 
    

Modified: 
    llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
    llvm/test/CodeGen/AMDGPU/amdgpu-inline.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
index eb2733ff0310..6728b07fb48b 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp
@@ -1149,7 +1149,9 @@ bool GCNTTIImpl::areInlineCompatible(const Function *Caller,
 
   // Hack to make compile times reasonable.
   if (InlineMaxBB && !Callee->hasFnAttribute(Attribute::InlineHint)) {
-    // Single BB does not increase total BB amount, thus subtract 1.
+    // Single BB does not increase total BB amount.
+    if (Callee->size() == 1)
+      return true;
     size_t BBSize = Caller->size() + Callee->size() - 1;
     return BBSize <= InlineMaxBB;
   }

diff  --git a/llvm/test/CodeGen/AMDGPU/amdgpu-inline.ll b/llvm/test/CodeGen/AMDGPU/amdgpu-inline.ll
index 9b0803bbcb99..7d02f393dc6e 100644
--- a/llvm/test/CodeGen/AMDGPU/amdgpu-inline.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdgpu-inline.ll
@@ -1,7 +1,8 @@
-; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -O3 -S -inline-threshold=1 < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-INL1 %s
-; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -O3 -S < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-INLDEF %s
-; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -passes='default<O3>' -S -inline-threshold=1 < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-INL1 %s
-; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -passes='default<O3>' -S < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-INLDEF %s
+; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -O3 -S -inline-threshold=1 < %s | FileCheck -check-prefixes=GCN,GCN-INL1,GCN-MAXBBDEF %s
+; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -O3 -S < %s | FileCheck -check-prefixes=GCN,GCN-INLDEF,GCN-MAXBBDEF %s
+; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -passes='default<O3>' -S -inline-threshold=1 < %s | FileCheck -check-prefixes=GCN,GCN-INL1,GCN-MAXBBDEF %s
+; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -passes='default<O3>' -S < %s | FileCheck -check-prefixes=GCN,GCN-INLDEF,GCN-MAXBBDEF %s
+; RUN: opt -mtriple=amdgcn--amdhsa -data-layout=A5 -passes='default<O3>' -S -amdgpu-inline-max-bb=1 < %s | FileCheck -check-prefixes=GCN,GCN-MAXBB1 %s
 
 define coldcc float @foo(float %x, float %y) {
 entry:
@@ -57,12 +58,14 @@ entry:
 }
 
 ; GCN: define amdgpu_kernel void @test_inliner(
-; GCN-INL1:   %c1 = tail call coldcc float @foo(
-; GCN-INLDEF: %cmp.i = fcmp ogt float %tmp2, 0.000000e+00
-; GCN:        %div.i{{[0-9]*}} = fdiv float 1.000000e+00, %c
-; GCN:        %div.i{{[0-9]*}} = fdiv float 2.000000e+00, %tmp1.i
-; GCN:        call void @foo_noinline(
-; GCN:        tail call float @_Z3sinf(
+; GCN-INL1:     %c1 = tail call coldcc float @foo(
+; GCN-INLDEF:   %cmp.i = fcmp ogt float %tmp2, 0.000000e+00
+; GCN-MAXBBDEF: %div.i{{[0-9]*}} = fdiv float 1.000000e+00, %c
+; GCN-MAXBBDEF: %div.i{{[0-9]*}} = fdiv float 2.000000e+00, %tmp1.i
+; GCN-MAXBB1:   call coldcc void @foo_private_ptr
+; GCN-MAXBB1:   call coldcc void @foo_private_ptr2
+; GCN:          call void @foo_noinline(
+; GCN:          tail call float @_Z3sinf(
 define amdgpu_kernel void @test_inliner(float addrspace(1)* nocapture %a, i32 %n) {
 entry:
   %pvt_arr = alloca [64 x float], align 4, addrspace(5)
@@ -95,7 +98,8 @@ entry:
 }
 
 ; GCN: define amdgpu_kernel void @test_inliner_multi_pvt_ptr(
-; GCN: %div.i{{[0-9]*}} = fdiv float 2.000000e+00, %tmp1.i
+; GCN-MAXBBDEF: %div.i{{[0-9]*}} = fdiv float 2.000000e+00, %tmp1.i
+; GCN-MAXBB1:   call coldcc void @foo_private_ptr2
 define amdgpu_kernel void @test_inliner_multi_pvt_ptr(float addrspace(1)* nocapture %a, i32 %n, float %v) {
 entry:
   %pvt_arr1 = alloca [32 x float], align 4, addrspace(5)
@@ -147,6 +151,24 @@ entry:
   ret void
 }
 
+; GCN: define amdgpu_kernel void @test_inliner_maxbb_singlebb(
+; GCN: tail call float @_Z3sinf
+define amdgpu_kernel void @test_inliner_maxbb_singlebb(float addrspace(1)* nocapture %a, i32 %n) {
+entry:
+  %cmp = icmp eq i32 %n, 1
+  br i1 %cmp, label %bb.1, label %bb.2
+  br label %bb.1
+
+bb.1:
+  store float 1.0, float* undef
+  br label %bb.2
+
+bb.2:
+  %c = call float @sin_wrapper(float 1.0)
+  store float %c, float addrspace(1)* %a
+  ret void
+}
+
 declare i32 @llvm.amdgcn.workitem.id.x() #1
 declare float @_Z3sinf(float) #1