[LoopInterchange] Handle confused dependence correctly (#140709)
This patch fixes the handling of a confused `Dependence` object. Such an object doesn’t contain any information about dependencies, so we must process it conservatively. However, it was converted into a direction vector like `[I I ... I]`. As a result, it was treated as if there are no loop-carried dependencies, which can lead to illegal loop exchanges. Fixes #140238
This commit is contained in:
parent
0c89cbb484
commit
1e5f7f64b0
@ -209,6 +209,14 @@ static bool populateDependencyMatrix(CharMatrix &DepMatrix, unsigned Level,
|
||||
Direction = '*';
|
||||
Dep.push_back(Direction);
|
||||
}
|
||||
|
||||
// If the Dependence object doesn't have any information, fill the
|
||||
// dependency vector with '*'.
|
||||
if (D->isConfused()) {
|
||||
assert(Dep.empty() && "Expected empty dependency vector");
|
||||
Dep.assign(Level, '*');
|
||||
}
|
||||
|
||||
while (Dep.size() != Level) {
|
||||
Dep.push_back('I');
|
||||
}
|
||||
|
41
llvm/test/Transforms/LoopInterchange/confused-dependence.ll
Normal file
41
llvm/test/Transforms/LoopInterchange/confused-dependence.ll
Normal file
@ -0,0 +1,41 @@
|
||||
; REQUIRES: asserts
|
||||
; RUN: opt < %s -passes=loop-interchange -verify-dom-info -verify-loop-info \
|
||||
; RUN: -disable-output -debug 2>&1 | FileCheck %s
|
||||
|
||||
;; In the following case, p0 and p1 may alias, so the direction vector must be [* *].
|
||||
;;
|
||||
;; void may_alias(float *p0, float *p1) {
|
||||
;; for (int i = 0; i < 4; i++)
|
||||
;; for (int j = 0; j < 4; j++)
|
||||
;; p0[4 * i + j] = p1[4 * j + i];
|
||||
;; }
|
||||
|
||||
; CHECK: Dependency matrix before interchange:
|
||||
; CHECK-NEXT: * *
|
||||
; CHECK-NEXT: Processing InnerLoopId = 1 and OuterLoopId = 0
|
||||
define void @may_alias(ptr %p0, ptr %p1) {
|
||||
entry:
|
||||
br label %for.i.header
|
||||
|
||||
for.i.header:
|
||||
%i = phi i32 [ 0, %entry ], [ %i.inc, %for.i.latch ]
|
||||
br label %for.j
|
||||
|
||||
for.j:
|
||||
%j = phi i32 [ 0, %for.i.header ], [ %j.inc, %for.j ]
|
||||
%idx.0 = getelementptr inbounds [4 x [4 x float]], ptr %p0, i32 0, i32 %j, i32 %i
|
||||
%idx.1 = getelementptr inbounds [4 x [4 x float]], ptr %p1, i32 0, i32 %i, i32 %j
|
||||
%0 = load float, ptr %idx.0, align 4
|
||||
store float %0, ptr %idx.1, align 4
|
||||
%j.inc = add nuw nsw i32 %j, 1
|
||||
%cmp.j = icmp slt i32 %j.inc, 4
|
||||
br i1 %cmp.j, label %for.j, label %for.i.latch
|
||||
|
||||
for.i.latch:
|
||||
%i.inc = add nuw nsw i32 %i, 1
|
||||
%cmp.i = icmp slt i32 %i.inc, 4
|
||||
br i1 %cmp.i, label %for.i.header, label %exit
|
||||
|
||||
exit:
|
||||
ret void
|
||||
}
|
@ -4,7 +4,7 @@
|
||||
|
||||
; Triply nested loop, should be able to do interchange three times
|
||||
; to get the ideal access pattern.
|
||||
; void f(int e[10][10][10], int f[10][10][10]) {
|
||||
; void f(int e[restrict 10][10][10], int f[restrict 10][10][10]) {
|
||||
; for (int a = 0; a < 10; a++) {
|
||||
; for (int b = 0; b < 10; b++) {
|
||||
; for (int c = 0; c < 10; c++) {
|
||||
@ -35,7 +35,7 @@
|
||||
; REMARKS-NEXT: Name: Interchanged
|
||||
; REMARKS-NEXT: Function: pr43326-triply-nested
|
||||
|
||||
define void @pr43326-triply-nested(ptr %e, ptr %f) {
|
||||
define void @pr43326-triply-nested(ptr noalias %e, ptr noalias %f) {
|
||||
entry:
|
||||
br label %for.outermost.header
|
||||
|
||||
|
@ -2,7 +2,7 @@
|
||||
; RUN: opt < %s -passes=loop-interchange -S -debug 2>&1 | FileCheck %s
|
||||
|
||||
; CHECK: Dependency matrix before interchange:
|
||||
; CHECK-NEXT: I I
|
||||
; CHECK-NEXT: * *
|
||||
; CHECK-NEXT: = *
|
||||
; CHECK-NEXT: < *
|
||||
; CHECK-NEXT: Processing InnerLoopId
|
||||
|
Loading…
x
Reference in New Issue
Block a user