[LV] Calculating explicit vector length using vector trip count #89580
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-llvm-transforms Author: Mel Chen (Mel-Chen) ChangesAlign with TODO: Return Full diff: https://github.com/llvm/llvm-project/pull/89580.diff 7 Files Affected:
diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index 5535cc55e93216..fa9cc2c21971d7 100644
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -2760,6 +2760,13 @@ InnerLoopVectorizer::getOrCreateVectorTripCount(BasicBlock *InsertBlock) {
IRBuilder<> Builder(InsertBlock->getTerminator());
Type *Ty = TC->getType();
+ // If the tail is to be folded by EVL, typically the original trip count can
+ // be used directly as the vector trip count.
+ // TODO: However, in cases where the epilogue is required, the last iteration
+ // needs to be left for execution by the epilogue.
+ if (Cost->foldTailWithEVL())
+ return VectorTripCount = TC;
+
// This is where we can make the step a runtime constant.
Value *Step = createStepForVF(Builder, Ty, VF, UF);
diff --git a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
index 1256e4d8fda50b..d48a9151e5a05e 100644
--- a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
+++ b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
@@ -1297,7 +1297,7 @@ void VPlanTransforms::addActiveLaneMask(
/// ...
/// %EVLPhi = EXPLICIT-VECTOR-LENGTH-BASED-IV-PHI [ %StartV, %vector.ph ],
/// [ %NextEVLIV, %vector.body ]
-/// %VPEVL = EXPLICIT-VECTOR-LENGTH %EVLPhi, original TC
+/// %VPEVL = EXPLICIT-VECTOR-LENGTH %EVLPhi, vector TC
/// ...
/// %NextEVLIV = add IVSize (cast i32 %VPEVVL to IVSize), %EVLPhi
/// ...
@@ -1321,7 +1321,7 @@ void VPlanTransforms::addExplicitVectorLength(VPlan &Plan) {
auto *EVLPhi = new VPEVLBasedIVPHIRecipe(StartV, DebugLoc());
EVLPhi->insertAfter(CanonicalIVPHI);
auto *VPEVL = new VPInstruction(VPInstruction::ExplicitVectorLength,
- {EVLPhi, Plan.getTripCount()});
+ {EVLPhi, &Plan.getVectorTripCount()});
VPEVL->insertBefore(*Header, Header->getFirstNonPhi());
auto *CanonicalIVIncrement =
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-gather-scatter.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-gather-scatter.ll
index 835ff375688173..aacae3de045828 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-gather-scatter.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-gather-scatter.ll
@@ -18,14 +18,6 @@ define void @gather_scatter(ptr noalias %in, ptr noalias %out, ptr noalias %inde
; IF-EVL-NEXT: [[TMP3:%.*]] = icmp ult i64 [[TMP0]], [[TMP2]]
; IF-EVL-NEXT: br i1 [[TMP3]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; IF-EVL: vector.ph:
-; IF-EVL-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
-; IF-EVL-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 2
-; IF-EVL-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
-; IF-EVL-NEXT: [[TMP7:%.*]] = mul i64 [[TMP6]], 2
-; IF-EVL-NEXT: [[TMP8:%.*]] = sub i64 [[TMP7]], 1
-; IF-EVL-NEXT: [[N_RND_UP:%.*]] = add i64 [[N]], [[TMP8]]
-; IF-EVL-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
-; IF-EVL-NEXT: [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
; IF-EVL-NEXT: [[TRIP_COUNT_MINUS_1:%.*]] = sub i64 [[N]], 1
; IF-EVL-NEXT: [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
; IF-EVL-NEXT: [[TMP10:%.*]] = mul i64 [[TMP9]], 2
@@ -58,12 +50,12 @@ define void @gather_scatter(ptr noalias %in, ptr noalias %out, ptr noalias %inde
; IF-EVL-NEXT: [[INDEX_EVL_NEXT]] = add i64 [[TMP23]], [[EVL_BASED_IV]]
; IF-EVL-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX1]], [[TMP10]]
; IF-EVL-NEXT: [[VEC_IND_NEXT]] = add <vscale x 2 x i64> [[VEC_IND]], [[DOTSPLAT]]
-; IF-EVL-NEXT: [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT: [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N]]
; IF-EVL-NEXT: br i1 [[TMP24]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; IF-EVL: middle.block:
; IF-EVL-NEXT: br i1 true, label [[FOR_END:%.*]], label [[SCALAR_PH]]
; IF-EVL: scalar.ph:
-; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; IF-EVL-NEXT: br label [[FOR_BODY:%.*]]
; IF-EVL: for.body:
; IF-EVL-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-iv32.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-iv32.ll
index d5ad99f5cff884..56ec351ac9559b 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-iv32.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-iv32.ll
@@ -18,14 +18,6 @@ define void @iv32(ptr noalias %a, ptr noalias %b, i32 %N) {
; IF-EVL-NEXT: [[TMP3:%.*]] = icmp ult i32 [[TMP19]], [[TMP2]]
; IF-EVL-NEXT: br i1 [[TMP3]], label [[SCALAR_PH:%.*]], label [[ENTRY:%.*]]
; IF-EVL: vector.ph:
-; IF-EVL-NEXT: [[TMP4:%.*]] = call i32 @llvm.vscale.i32()
-; IF-EVL-NEXT: [[TMP5:%.*]] = mul i32 [[TMP4]], 4
-; IF-EVL-NEXT: [[TMP6:%.*]] = call i32 @llvm.vscale.i32()
-; IF-EVL-NEXT: [[TMP7:%.*]] = mul i32 [[TMP6]], 4
-; IF-EVL-NEXT: [[TMP8:%.*]] = sub i32 [[TMP7]], 1
-; IF-EVL-NEXT: [[N_RND_UP:%.*]] = add i32 [[N]], [[TMP8]]
-; IF-EVL-NEXT: [[N_MOD_VF:%.*]] = urem i32 [[N_RND_UP]], [[TMP5]]
-; IF-EVL-NEXT: [[N_VEC:%.*]] = sub i32 [[N_RND_UP]], [[N_MOD_VF]]
; IF-EVL-NEXT: [[TMP9:%.*]] = call i32 @llvm.vscale.i32()
; IF-EVL-NEXT: [[TMP10:%.*]] = mul i32 [[TMP9]], 4
; IF-EVL-NEXT: br label [[FOR_BODY:%.*]]
@@ -43,12 +35,12 @@ define void @iv32(ptr noalias %a, ptr noalias %b, i32 %N) {
; IF-EVL-NEXT: call void @llvm.vp.store.nxv4i32.p0(<vscale x 4 x i32> [[VP_OP_LOAD]], ptr align 4 [[TMP17]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP12]])
; IF-EVL-NEXT: [[INDEX_EVL_NEXT]] = add i32 [[TMP12]], [[EVL_BASED_IV]]
; IF-EVL-NEXT: [[IV_NEXT]] = add i32 [[IV]], [[TMP10]]
-; IF-EVL-NEXT: [[TMP18:%.*]] = icmp eq i32 [[IV_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT: [[TMP18:%.*]] = icmp eq i32 [[IV_NEXT]], [[N]]
; IF-EVL-NEXT: br i1 [[TMP18]], label [[MIDDLE_BLOCK:%.*]], label [[FOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; IF-EVL: middle.block:
; IF-EVL-NEXT: br i1 true, label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH]]
; IF-EVL: scalar.ph:
-; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY1:%.*]] ]
+; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY1:%.*]] ]
; IF-EVL-NEXT: br label [[FOR_BODY1:%.*]]
; IF-EVL: for.body:
; IF-EVL-NEXT: [[IV1:%.*]] = phi i32 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT1:%.*]], [[FOR_BODY1]] ]
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll
index 203d0c977074e9..ed40af04fd12b5 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-force-tail-with-evl-masked-loadstore.ll
@@ -18,14 +18,6 @@ define void @masked_loadstore(ptr noalias %a, ptr noalias %b, i64 %n) {
; IF-EVL-NEXT: [[TMP3:%.*]] = icmp ult i64 [[TMP0]], [[TMP2]]
; IF-EVL-NEXT: br i1 [[TMP3]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; IF-EVL: vector.ph:
-; IF-EVL-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
-; IF-EVL-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 4
-; IF-EVL-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
-; IF-EVL-NEXT: [[TMP7:%.*]] = mul i64 [[TMP6]], 4
-; IF-EVL-NEXT: [[TMP8:%.*]] = sub i64 [[TMP7]], 1
-; IF-EVL-NEXT: [[N_RND_UP:%.*]] = add i64 [[N]], [[TMP8]]
-; IF-EVL-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
-; IF-EVL-NEXT: [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
; IF-EVL-NEXT: [[TRIP_COUNT_MINUS_1:%.*]] = sub i64 [[N]], 1
; IF-EVL-NEXT: [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
; IF-EVL-NEXT: [[TMP10:%.*]] = mul i64 [[TMP9]], 4
@@ -57,12 +49,12 @@ define void @masked_loadstore(ptr noalias %a, ptr noalias %b, i64 %n) {
; IF-EVL-NEXT: [[TMP24:%.*]] = zext i32 [[TMP12]] to i64
; IF-EVL-NEXT: [[INDEX_EVL_NEXT]] = add i64 [[TMP24]], [[EVL_BASED_IV]]
; IF-EVL-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
-; IF-EVL-NEXT: [[TMP25:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT: [[TMP25:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N]]
; IF-EVL-NEXT: br i1 [[TMP25]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; IF-EVL: middle.block:
; IF-EVL-NEXT: br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]]
; IF-EVL: scalar.ph:
-; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; IF-EVL-NEXT: br label [[FOR_BODY:%.*]]
; IF-EVL: for.body:
; IF-EVL-NEXT: [[I_011:%.*]] = phi i64 [ [[INC:%.*]], [[FOR_INC:%.*]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll
index c69bb17f698aaf..3b97690ae6db44 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vectorize-vp-intrinsics.ll
@@ -18,14 +18,6 @@ define void @foo(ptr noalias %a, ptr noalias %b, ptr noalias %c, i64 %N) {
; IF-EVL-NEXT: [[TMP3:%.*]] = icmp ult i64 [[TMP0]], [[TMP2]]
; IF-EVL-NEXT: br i1 [[TMP3]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; IF-EVL: vector.ph:
-; IF-EVL-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
-; IF-EVL-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 4
-; IF-EVL-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
-; IF-EVL-NEXT: [[TMP7:%.*]] = mul i64 [[TMP6]], 4
-; IF-EVL-NEXT: [[TMP8:%.*]] = sub i64 [[TMP7]], 1
-; IF-EVL-NEXT: [[N_RND_UP:%.*]] = add i64 [[N]], [[TMP8]]
-; IF-EVL-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP5]]
-; IF-EVL-NEXT: [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
; IF-EVL-NEXT: [[TMP9:%.*]] = call i64 @llvm.vscale.i64()
; IF-EVL-NEXT: [[TMP10:%.*]] = mul i64 [[TMP9]], 4
; IF-EVL-NEXT: br label [[VECTOR_BODY:%.*]]
@@ -48,12 +40,12 @@ define void @foo(ptr noalias %a, ptr noalias %b, ptr noalias %c, i64 %N) {
; IF-EVL-NEXT: [[TMP21:%.*]] = zext i32 [[TMP12]] to i64
; IF-EVL-NEXT: [[INDEX_EVL_NEXT]] = add i64 [[TMP21]], [[EVL_BASED_IV]]
; IF-EVL-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP10]]
-; IF-EVL-NEXT: [[TMP22:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; IF-EVL-NEXT: [[TMP22:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N]]
; IF-EVL-NEXT: br i1 [[TMP22]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; IF-EVL: middle.block:
; IF-EVL-NEXT: br i1 true, label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH]]
; IF-EVL: scalar.ph:
-; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; IF-EVL-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; IF-EVL-NEXT: br label [[FOR_BODY:%.*]]
; IF-EVL: for.body:
; IF-EVL-NEXT: [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[FOR_BODY]] ]
diff --git a/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll b/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll
index 72b881bd44c768..ce8deed6c92cd5 100644
--- a/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll
+++ b/llvm/test/Transforms/LoopVectorize/RISCV/vplan-vp-intrinsics.ll
@@ -23,7 +23,7 @@ define void @foo(ptr noalias %a, ptr noalias %b, ptr noalias %c, i64 %N) {
; IF-EVL-NEXT: vector.body:
; IF-EVL-NEXT: EMIT vp<[[IV:%[0-9]+]]> = CANONICAL-INDUCTION
; IF-EVL-NEXT: EXPLICIT-VECTOR-LENGTH-BASED-IV-PHI vp<[[EVL_PHI:%[0-9]+]]> = phi ir<0>, vp<[[IV_NEXT:%[0-9]+]]>
-; IF-EVL-NEXT: EMIT vp<[[EVL:%.+]]> = EXPLICIT-VECTOR-LENGTH vp<[[EVL_PHI]]>, ir<%N>
+; IF-EVL-NEXT: EMIT vp<[[EVL:%.+]]> = EXPLICIT-VECTOR-LENGTH vp<[[EVL_PHI]]>, vp<[[VTC]]>
; IF-EVL-NEXT: vp<[[ST:%[0-9]+]]> = SCALAR-STEPS vp<[[EVL_PHI]]>, ir<1>
; IF-EVL-NEXT: CLONE ir<[[GEP1:%.+]]> = getelementptr inbounds ir<%b>, vp<[[ST]]>
; IF-EVL-NEXT: vp<[[PTR1:%[0-9]+]]> = vector-pointer ir<[[GEP1]]>
|
fhahn
reviewed
Apr 22, 2024
| @@ -58,12 +50,12 @@ define void @gather_scatter(ptr noalias %in, ptr noalias %out, ptr noalias %inde | |||
| ; IF-EVL-NEXT: [[INDEX_EVL_NEXT]] = add i64 [[TMP23]], [[EVL_BASED_IV]] | |||
| ; IF-EVL-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX1]], [[TMP10]] | |||
| ; IF-EVL-NEXT: [[VEC_IND_NEXT]] = add <vscale x 2 x i64> [[VEC_IND]], [[DOTSPLAT]] | |||
| ; IF-EVL-NEXT: [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]] | |||
| ; IF-EVL-NEXT: [[TMP24:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N]] | |||
There was a problem hiding this comment.
I'm not sure how this is correct if %N isn't a multiple of the VF? INDEX_NEXT is incremented by 2 * vscale;
There was a problem hiding this comment.
My bad, you are right.
I think I should first use index.evl.next as the condition for branch-on-count instead of vec.ind.next, or directly reuse the canonical IV as evl.based.iv and change its step to evl.
Do you have any suggestions or preferences?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Align with
branch-on-count, calculate explicit vector length using the vector trip count.TODO: Return
original trip count - 1as vector trip count when requiring epilogue.