RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-19 11:02:59 +02:00
Code Issues Projects Releases Wiki Activity

[SLSR] consider &B[S << i] as &B[(1 << i) * S]

Browse Source 
Summary: This reduces handling &B[(1 << i) * s] to handling &B[i * S].

Test Plan: slsr-gep.ll

Reviewers: meheff

Subscribers: sanjoy, llvm-commits

Differential Revision: http://reviews.llvm.org/D8837

llvm-svn: 234180
This commit is contained in:
Jingyue Wu 2015-04-06 17:15:48 +00:00
parent e8ed19bff1
commit cee19c8e7e
2 changed files with 12 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
lib/Transforms/Scalar/StraightLineStrengthReduce.cpp
View File

@ -353,8 +353,6 @@ void StraightLineStrengthReduce::factorArrayIndex(Value *ArrayIdx,
ArrayIdx, ElementSize, GEP);
Value *LHS = nullptr;
ConstantInt *RHS = nullptr;
// TODO: handle shl. e.g., we could treat (S << 2) as (S * 4).
//
// One alternative is matching the SCEV of ArrayIdx instead of ArrayIdx
// itself. This would allow us to handle the shl case for free. However,
// matching SCEVs has two issues:
@ -370,6 +368,13 @@ void StraightLineStrengthReduce::factorArrayIndex(Value *ArrayIdx,
// SLSR is currently unsafe if i * S may overflow.
// GEP = Base + sext(LHS *nsw RHS) * ElementSize
allocateCandidateAndFindBasisForGEP(Base, RHS, LHS, ElementSize, GEP);
} else if (match(ArrayIdx, m_NSWShl(m_Value(LHS), m_ConstantInt(RHS)))) {
// GEP = Base + sext(LHS <<nsw RHS) * ElementSize
// = Base + sext(LHS *nsw (1 << RHS)) * ElementSize
APInt One(RHS->getBitWidth(), 1);
ConstantInt *PowerOf2 =
ConstantInt::get(RHS->getContext(), One << RHS->getValue());
allocateCandidateAndFindBasisForGEP(Base, PowerOf2, LHS, ElementSize, GEP);
}
}

10
test/Transforms/StraightLineStrengthReduce/slsr-gep.ll
View File

@ -14,7 +14,7 @@ define i32 @slsr_gep(i32* %input, i64 %s) {
%v1 = load i32, i32* %p1
; v2 = input[s * 2];
%s2 = mul nsw i64 %s, 2
%s2 = shl nsw i64 %s, 1
%p2 = getelementptr inbounds i32, i32* %input, i64 %s2
; CHECK: %p2 = getelementptr inbounds i32, i32* %p1, i64 %s
%v2 = load i32, i32* %p2
@ -38,7 +38,7 @@ define i32 @slsr_gep_sext(i32* %input, i32 %s) {
%v1 = load i32, i32* %p1
; v2 = input[(long)(s * 2)];
%s2 = mul nsw i32 %s, 2
%s2 = shl nsw i32 %s, 1
%t2 = sext i32 %s2 to i64
%p2 = getelementptr inbounds i32, i32* %input, i64 %t2
; CHECK: %p2 = getelementptr inbounds i32, i32* %p1, i64 %t
@ -57,7 +57,7 @@ define i32 @slsr_gep_2d([10 x [5 x i32]]* %input, i64 %s, i64 %t) {
%v0 = load i32, i32* %p0
; v1 = input[s * 2][t];
%s2 = mul nsw i64 %s, 2
%s2 = shl nsw i64 %s, 1
; CHECK: [[BUMP:%[a-zA-Z0-9]+]] = mul i64 %s, 5
%p1 = getelementptr inbounds [10 x [5 x i32]], [10 x [5 x i32]]* %input, i64 0, i64 %s2, i64 %t
; CHECK: %p1 = getelementptr inbounds i32, i32* %p0, i64 [[BUMP]]
@ -90,7 +90,7 @@ define i64 @slsr_gep_uglygep([10 x [5 x %struct.S]]* %input, i64 %s, i64 %t) {
%v0 = load i64, i64* %p0
; v1 = input[s * 2][t].f1;
%s2 = mul nsw i64 %s, 2
%s2 = shl nsw i64 %s, 1
; CHECK: [[BUMP:%[a-zA-Z0-9]+]] = mul i64 %s, 60
%p1 = getelementptr inbounds [10 x [5 x %struct.S]], [10 x [5 x %struct.S]]* %input, i64 0, i64 %s2, i64 %t, i32 0
; CHECK: getelementptr inbounds i8, i8* %{{[0-9]+}}, i64 [[BUMP]]
@ -121,7 +121,7 @@ define i32 @slsr_out_of_bounds_gep(i32* %input, i32 %s) {
%v1 = load i32, i32* %p1
; v2 = input[(long)(s * 2)];
%s2 = mul nsw i32 %s, 2
%s2 = shl nsw i32 %s, 1
%t2 = sext i32 %s2 to i64
%p2 = getelementptr i32, i32* %input, i64 %t2
; CHECK: %p2 = getelementptr i32, i32* %p1, i64 %t