[DA] Set Distance to zero when Direction is EQ (#147966)

A Dependence object has two entries: Distance and Direction. The former
represents the distance of the dependence, while the latter
characterizes the distance by whether the value of it is positive,
negative, or zero (especially, zero is represented by EQ in DA). So it
is expected that the Distance equals zero iff the Direction is EQ.
However, this condition was not satisfied in some cases.
This patch adds a logic to set the Distance to zero if the Direction is
EQ. Although it is ideal that the Distance is updated to zero
simultaneously when the Direction is set to EQ, achieving it would
require changing the entire code in DA.
This commit is contained in:
Ryotaro Kasuga 2025-07-11 09:22:24 +09:00 committed by GitHub
parent c0b82df5f3
commit 7d510b7f21
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 49 additions and 13 deletions

View File

@ -187,6 +187,20 @@ static void dumpExampleDependence(raw_ostream &OS, DependenceInfo *DA,
OS << " da analyze - ";
if (auto D = DA->depends(&*SrcI, &*DstI,
/*UnderRuntimeAssumptions=*/true)) {
#ifndef NDEBUG
// Verify that the distance being zero is equivalent to the
// direction being EQ.
for (unsigned Level = 1; Level <= D->getLevels(); Level++) {
const SCEV *Distance = D->getDistance(Level);
bool IsDistanceZero = Distance && Distance->isZero();
bool IsDirectionEQ =
D->getDirection(Level) == Dependence::DVEntry::EQ;
assert(IsDistanceZero == IsDirectionEQ &&
"Inconsistent distance and direction.");
}
#endif
// Normalize negative direction vectors if required by clients.
if (NormalizeResults && D->normalize(&SE))
OS << "normalized - ";
@ -3991,6 +4005,28 @@ DependenceInfo::depends(Instruction *Src, Instruction *Dst,
if (CompleteLoops[II])
Result.DV[II - 1].Scalar = false;
// Set the distance to zero if the direction is EQ.
// TODO: Ideally, the distance should be set to 0 immediately simultaneously
// with the corresponding direction being set to EQ.
for (unsigned II = 1; II <= Result.getLevels(); ++II) {
if (Result.getDirection(II) == Dependence::DVEntry::EQ) {
if (Result.DV[II - 1].Distance == nullptr)
Result.DV[II - 1].Distance = SE->getZero(SrcSCEV->getType());
else
assert(Result.DV[II - 1].Distance->isZero() &&
"Inconsistency between distance and direction");
}
#ifndef NDEBUG
// Check that the converse (i.e., if the distance is zero, then the
// direction is EQ) holds.
const SCEV *Distance = Result.getDistance(II);
if (Distance && Distance->isZero())
assert(Result.getDirection(II) == Dependence::DVEntry::EQ &&
"Distance is zero, but direction is not EQ");
#endif
}
if (PossiblyLoopIndependent) {
// Make sure the LoopIndependent flag is set correctly.
// All directions must include equal, otherwise no

View File

@ -802,7 +802,7 @@ define void @banerjee9(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; CHECK-NEXT: da analyze - output [* *]!
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %1 = load i64, ptr %arrayidx7, align 8
; CHECK-NEXT: da analyze - flow [<= =|<]!
; CHECK-NEXT: da analyze - flow [<= 0|<]!
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %1, ptr %B.addr.11, align 8
; CHECK-NEXT: da analyze - confused!
; CHECK-NEXT: Src: %1 = load i64, ptr %arrayidx7, align 8 --> Dst: %1 = load i64, ptr %arrayidx7, align 8
@ -816,7 +816,7 @@ define void @banerjee9(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; NORMALIZE-NEXT: da analyze - output [* *]!
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %1 = load i64, ptr %arrayidx7, align 8
; NORMALIZE-NEXT: da analyze - flow [<= =|<]!
; NORMALIZE-NEXT: da analyze - flow [<= 0|<]!
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %1, ptr %B.addr.11, align 8
; NORMALIZE-NEXT: da analyze - confused!
; NORMALIZE-NEXT: Src: %1 = load i64, ptr %arrayidx7, align 8 --> Dst: %1 = load i64, ptr %arrayidx7, align 8
@ -830,7 +830,7 @@ define void @banerjee9(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; DELIN-NEXT: da analyze - output [* *]!
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %1 = load i64, ptr %arrayidx7, align 8
; DELIN-NEXT: da analyze - flow [<= =|<]!
; DELIN-NEXT: da analyze - flow [<= 0|<]!
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %1, ptr %B.addr.11, align 8
; DELIN-NEXT: da analyze - confused!
; DELIN-NEXT: Src: %1 = load i64, ptr %arrayidx7, align 8 --> Dst: %1 = load i64, ptr %arrayidx7, align 8
@ -888,7 +888,7 @@ define void @banerjee10(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; CHECK-NEXT: da analyze - none!
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %1 = load i64, ptr %arrayidx6, align 8
; CHECK-NEXT: da analyze - flow [<> =]!
; CHECK-NEXT: da analyze - flow [<> 0]!
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %1, ptr %B.addr.11, align 8
; CHECK-NEXT: da analyze - confused!
; CHECK-NEXT: Src: %1 = load i64, ptr %arrayidx6, align 8 --> Dst: %1 = load i64, ptr %arrayidx6, align 8
@ -902,7 +902,7 @@ define void @banerjee10(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; NORMALIZE-NEXT: da analyze - none!
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %1 = load i64, ptr %arrayidx6, align 8
; NORMALIZE-NEXT: da analyze - flow [<> =]!
; NORMALIZE-NEXT: da analyze - flow [<> 0]!
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %1, ptr %B.addr.11, align 8
; NORMALIZE-NEXT: da analyze - confused!
; NORMALIZE-NEXT: Src: %1 = load i64, ptr %arrayidx6, align 8 --> Dst: %1 = load i64, ptr %arrayidx6, align 8
@ -916,7 +916,7 @@ define void @banerjee10(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; DELIN-NEXT: da analyze - none!
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %1 = load i64, ptr %arrayidx6, align 8
; DELIN-NEXT: da analyze - flow [<> =]!
; DELIN-NEXT: da analyze - flow [<> 0]!
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %1, ptr %B.addr.11, align 8
; DELIN-NEXT: da analyze - confused!
; DELIN-NEXT: Src: %1 = load i64, ptr %arrayidx6, align 8 --> Dst: %1 = load i64, ptr %arrayidx6, align 8
@ -1058,7 +1058,7 @@ define void @banerjee12(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; CHECK-NEXT: da analyze - none!
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %0 = load i64, ptr %arrayidx6, align 8
; CHECK-NEXT: da analyze - flow [= <>]!
; CHECK-NEXT: da analyze - flow [0 <>]!
; CHECK-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %0, ptr %B.addr.11, align 8
; CHECK-NEXT: da analyze - confused!
; CHECK-NEXT: Src: %0 = load i64, ptr %arrayidx6, align 8 --> Dst: %0 = load i64, ptr %arrayidx6, align 8
@ -1072,7 +1072,7 @@ define void @banerjee12(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; NORMALIZE-NEXT: da analyze - none!
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %0 = load i64, ptr %arrayidx6, align 8
; NORMALIZE-NEXT: da analyze - flow [= <>]!
; NORMALIZE-NEXT: da analyze - flow [0 <>]!
; NORMALIZE-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %0, ptr %B.addr.11, align 8
; NORMALIZE-NEXT: da analyze - confused!
; NORMALIZE-NEXT: Src: %0 = load i64, ptr %arrayidx6, align 8 --> Dst: %0 = load i64, ptr %arrayidx6, align 8
@ -1086,7 +1086,7 @@ define void @banerjee12(ptr %A, ptr %B, i64 %m, i64 %n) nounwind uwtable ssp {
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 0, ptr %arrayidx, align 8
; DELIN-NEXT: da analyze - none!
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: %0 = load i64, ptr %arrayidx6, align 8
; DELIN-NEXT: da analyze - flow [= <>]!
; DELIN-NEXT: da analyze - flow [0 <>]!
; DELIN-NEXT: Src: store i64 0, ptr %arrayidx, align 8 --> Dst: store i64 %0, ptr %B.addr.11, align 8
; DELIN-NEXT: da analyze - confused!
; DELIN-NEXT: Src: %0 = load i64, ptr %arrayidx6, align 8 --> Dst: %0 = load i64, ptr %arrayidx6, align 8

View File

@ -285,7 +285,7 @@ define void @couple6(ptr %A, ptr %B, i32 %n) nounwind uwtable ssp {
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx1, align 4 --> Dst: store i32 %conv, ptr %arrayidx1, align 4
; CHECK-NEXT: da analyze - none!
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx1, align 4 --> Dst: %0 = load i32, ptr %arrayidx3, align 4
; CHECK-NEXT: da analyze - flow [=|<]!
; CHECK-NEXT: da analyze - flow [0|<]!
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx1, align 4 --> Dst: store i32 %0, ptr %B.addr.01, align 4
; CHECK-NEXT: da analyze - confused!
; CHECK-NEXT: Src: %0 = load i32, ptr %arrayidx3, align 4 --> Dst: %0 = load i32, ptr %arrayidx3, align 4
@ -503,7 +503,7 @@ define void @couple11(ptr %A, ptr %B, i32 %n) nounwind uwtable ssp {
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx2, align 4 --> Dst: store i32 %conv, ptr %arrayidx2, align 4
; CHECK-NEXT: da analyze - none!
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx2, align 4 --> Dst: %0 = load i32, ptr %arrayidx4, align 4
; CHECK-NEXT: da analyze - flow [=|<] splitable!
; CHECK-NEXT: da analyze - flow [0|<] splitable!
; CHECK-NEXT: da analyze - split level = 1, iteration = 9!
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx2, align 4 --> Dst: store i32 %0, ptr %B.addr.01, align 4
; CHECK-NEXT: da analyze - confused!
@ -636,7 +636,7 @@ define void @couple14(ptr %A, ptr %B, i32 %n) nounwind uwtable ssp {
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx3, align 4 --> Dst: store i32 %conv, ptr %arrayidx3, align 4
; CHECK-NEXT: da analyze - none!
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx3, align 4 --> Dst: %0 = load i32, ptr %arrayidx6, align 4
; CHECK-NEXT: da analyze - flow [=|<]!
; CHECK-NEXT: da analyze - flow [0|<]!
; CHECK-NEXT: Src: store i32 %conv, ptr %arrayidx3, align 4 --> Dst: store i32 %0, ptr %B.addr.01, align 4
; CHECK-NEXT: da analyze - confused!
; CHECK-NEXT: Src: %0 = load i32, ptr %arrayidx6, align 4 --> Dst: %0 = load i32, ptr %arrayidx6, align 4

View File

@ -18,7 +18,7 @@ define void @i32_subscript(ptr %a, ptr %b) {
; CHECK-NEXT: Src: %0 = load i32, ptr %a.addr, align 4 --> Dst: %0 = load i32, ptr %a.addr, align 4
; CHECK-NEXT: da analyze - none!
; CHECK-NEXT: Src: %0 = load i32, ptr %a.addr, align 4 --> Dst: store i32 %1, ptr %a.addr.2, align 4
; CHECK-NEXT: da analyze - anti [=|<]!
; CHECK-NEXT: da analyze - anti [0|<]!
; CHECK-NEXT: Src: store i32 %1, ptr %a.addr.2, align 4 --> Dst: store i32 %1, ptr %a.addr.2, align 4
; CHECK-NEXT: da analyze - none!
;