1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-22 18:54:02 +01:00

Add handling of !invariant.load to PropagateMetadata.

Summary:
This will let e.g. the load/store vectorizer propagate this metadata
appropriately.

Reviewers: arsenm

Subscribers: tra, jholewinski, hfinkel, mzolotukhin

Differential Revision: https://reviews.llvm.org/D23479

llvm-svn: 281153
This commit is contained in:
Justin Lebar 2016-09-11 01:39:08 +00:00
parent 1c4f697044
commit 2c71fc8bb7
2 changed files with 23 additions and 6 deletions

View File

@ -454,9 +454,10 @@ Instruction *llvm::propagateMetadata(Instruction *Inst, ArrayRef<Value *> VL) {
SmallVector<std::pair<unsigned, MDNode *>, 4> Metadata;
I0->getAllMetadataOtherThanDebugLoc(Metadata);
for (auto Kind : { LLVMContext::MD_tbaa, LLVMContext::MD_alias_scope,
LLVMContext::MD_noalias, LLVMContext::MD_fpmath,
LLVMContext::MD_nontemporal }) {
for (auto Kind :
{LLVMContext::MD_tbaa, LLVMContext::MD_alias_scope,
LLVMContext::MD_noalias, LLVMContext::MD_fpmath,
LLVMContext::MD_nontemporal, LLVMContext::MD_invariant_load}) {
MDNode *MD = I0->getMetadata(Kind);
for (int J = 1, E = VL.size(); MD && J != E; ++J) {
@ -469,13 +470,12 @@ Instruction *llvm::propagateMetadata(Instruction *Inst, ArrayRef<Value *> VL) {
case LLVMContext::MD_alias_scope:
MD = MDNode::getMostGenericAliasScope(MD, IMD);
break;
case LLVMContext::MD_noalias:
MD = MDNode::intersect(MD, IMD);
break;
case LLVMContext::MD_fpmath:
MD = MDNode::getMostGenericFPMath(MD, IMD);
break;
case LLVMContext::MD_noalias:
case LLVMContext::MD_nontemporal:
case LLVMContext::MD_invariant_load:
MD = MDNode::intersect(MD, IMD);
break;
default:

View File

@ -0,0 +1,17 @@
; RUN: opt -load-store-vectorizer -march=nvptx64 -mcpu=sm_35 -S < %s | FileCheck %s
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v16:16:16-v32:32:32-v64:64:64-v128:128:128-n16:32:64"
target triple = "nvptx64-nvidia-cuda"
; CHECK-LABEL: @foo
define i32 @foo(i32* %ptr) {
%ptr1 = getelementptr i32, i32* %ptr, i32 1
%p1 = addrspacecast i32* %ptr1 to i32 addrspace(1)*
; CHECK: load <2 x i32>, <2 x i32>* %{{[0-9]+}}, align 8, !invariant.load !0
%v0 = load i32, i32* %ptr, align 8, !invariant.load !0
%v1 = load i32, i32* %ptr1, align 4, !invariant.load !0
%sum = add i32 %v0, %v1
ret i32 %sum
}
!0 = !{}