Skip to content

Commit

Permalink
[LV] Add support for uniform parameters on vectorized function varian…
Browse files Browse the repository at this point in the history 
…ts (#72891)

Parameters marked as uniform take a scalar value, assuming the value is
invariant in the scalar loop.
  • Loading branch information
@huntergr-arm
huntergr-arm committed Nov 28, 2023
1 parent af4d8d5 commit 104b7c6
Show file tree
Hide file tree
Showing 3 changed files with 82 additions and 24 deletions.
8 changes: 8 additions & 0 deletions llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7027,6 +7027,14 @@ void LoopVectorizationCostModel::setVectorizedCallDecision(ElementCount VF) {
switch (Param.ParamKind) {
case VFParamKind::Vector:
break;
case VFParamKind::OMP_Uniform: {
Value *ScalarParam = CI->getArgOperand(Param.ParamPos);
// Make sure the scalar parameter in the loop is invariant.
if (!PSE.getSE()->isLoopInvariant(PSE.getSCEV(ScalarParam),
TheLoop))
ParamsOk = false;
break;
}
case VFParamKind::GlobalPredicate:
UsesMask = true;
break;
Expand Down
81 changes: 68 additions & 13 deletions llvm/test/Transforms/LoopVectorize/AArch64/uniform-args-call-variants.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,17 +9,25 @@ define void @test_uniform(ptr noalias %dst, ptr readonly %src, i64 %uniform , i6
; CHECK-LABEL: define void @test_uniform
; CHECK-SAME: (ptr noalias [[DST:%.*]], ptr readonly [[SRC:%.*]], i64 [[UNIFORM:%.*]], i64 [[N:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[GEPSRC:%.*]] = getelementptr double, ptr [[SRC]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[DATA:%.*]] = load double, ptr [[GEPSRC]], align 8
; CHECK-NEXT: [[CALL:%.*]] = call double @foo(double [[DATA]], i64 [[UNIFORM]]) #[[ATTR1:[0-9]+]]
; CHECK-NEXT: [[GEPDST:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store double [[CALL]], ptr [[GEPDST]], align 8
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[N]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP:%.*]], label [[FOR_BODY]]
; CHECK-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT: [[TMP1:%.*]] = shl i64 [[TMP0]], 1
; CHECK-NEXT: [[TMP2:%.*]] = call i64 @llvm.usub.sat.i64(i64 [[N]], i64 [[TMP1]])
; CHECK-NEXT: [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 [[N]])
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr double, ptr [[SRC]], i64 [[INDEX]]
; CHECK-NEXT: [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x double> @llvm.masked.load.nxv2f64.p0(ptr [[TMP3]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x double> poison)
; CHECK-NEXT: [[TMP4:%.*]] = call <vscale x 2 x double> @foo_uniform(<vscale x 2 x double> [[WIDE_MASKED_LOAD]], i64 [[UNIFORM]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: call void @llvm.masked.store.nxv2f64.p0(<vscale x 2 x double> [[TMP4]], ptr [[TMP5]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT: [[TMP7:%.*]] = shl i64 [[TMP6]], 1
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP7]]
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX]], i64 [[TMP2]])
; CHECK-NEXT: [[TMP8:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i64 0
; CHECK-NEXT: br i1 [[TMP8]], label [[VECTOR_BODY]], label [[FOR_COND_CLEANUP:%.*]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
;
Expand All @@ -41,6 +49,50 @@ for.cond.cleanup:
ret void
}

define void @test_uniform_smaller_scalar(ptr noalias %dst, ptr readonly %src, i32 %uniform , i64 %n) #0 {
; CHECK-LABEL: define void @test_uniform_smaller_scalar
; CHECK-SAME: (ptr noalias [[DST:%.*]], ptr readonly [[SRC:%.*]], i32 [[UNIFORM:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT: [[TMP1:%.*]] = shl i64 [[TMP0]], 1
; CHECK-NEXT: [[TMP2:%.*]] = call i64 @llvm.usub.sat.i64(i64 [[N]], i64 [[TMP1]])
; CHECK-NEXT: [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 [[N]])
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr double, ptr [[SRC]], i64 [[INDEX]]
; CHECK-NEXT: [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x double> @llvm.masked.load.nxv2f64.p0(ptr [[TMP3]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x double> poison)
; CHECK-NEXT: [[TMP4:%.*]] = call <vscale x 2 x double> @bar_uniform(<vscale x 2 x double> [[WIDE_MASKED_LOAD]], i32 [[UNIFORM]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: call void @llvm.masked.store.nxv2f64.p0(<vscale x 2 x double> [[TMP4]], ptr [[TMP5]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[TMP6:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT: [[TMP7:%.*]] = shl i64 [[TMP6]], 1
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP7]]
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX]], i64 [[TMP2]])
; CHECK-NEXT: [[TMP8:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i64 0
; CHECK-NEXT: br i1 [[TMP8]], label [[VECTOR_BODY]], label [[FOR_COND_CLEANUP:%.*]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
;
entry:
br label %for.body

for.body:
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%gepsrc = getelementptr double, ptr %src, i64 %indvars.iv
%data = load double, ptr %gepsrc, align 8
%call = call double @bar(double %data, i32 %uniform) #2
%gepdst = getelementptr inbounds double, ptr %dst, i64 %indvars.iv
store double %call, ptr %gepdst
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, %n
br i1 %exitcond, label %for.cond.cleanup, label %for.body

for.cond.cleanup:
ret void
}

; If the parameter is not uniform, then we can't use the vector variant.
define void @test_uniform_not_invariant(ptr noalias %dst, ptr readonly %src, i64 %n) #0 {
; CHECK-LABEL: define void @test_uniform_not_invariant
Expand All @@ -51,7 +103,7 @@ define void @test_uniform_not_invariant(ptr noalias %dst, ptr readonly %src, i64
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[GEPSRC:%.*]] = getelementptr double, ptr [[SRC]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[DATA:%.*]] = load double, ptr [[GEPSRC]], align 8
; CHECK-NEXT: [[CALL:%.*]] = call double @foo(double [[DATA]], i64 [[INDVARS_IV]]) #[[ATTR1]]
; CHECK-NEXT: [[CALL:%.*]] = call double @foo(double [[DATA]], i64 [[INDVARS_IV]]) #[[ATTR5:[0-9]+]]
; CHECK-NEXT: [[GEPDST:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store double [[CALL]], ptr [[GEPDST]], align 8
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
Expand Down Expand Up @@ -80,11 +132,14 @@ for.cond.cleanup:

; Scalar functions
declare double @foo(double, i64)
declare double @bar(double, i32)

; Vector variants
declare <vscale x 2 x double> @foo_uniform(<vscale x 2 x double>, i64, <vscale x 2 x i1>)
declare <vscale x 2 x double> @bar_uniform(<vscale x 2 x double>, i32, <vscale x 2 x i1>)

attributes #0 = { "target-features"="+sve" }

; Mappings
attributes #1 = { nounwind "vector-function-abi-variant"="_ZGV_LLVM_Mxvu_foo(foo_uniform)" }
attributes #1 = { nounwind "vector-function-abi-variant"="_ZGVsMxvu_foo(foo_uniform)" }
attributes #2 = { nounwind "vector-function-abi-variant"="_ZGVsMxvu_bar(bar_uniform)" }
17 changes: 6 additions & 11 deletions llvm/test/Transforms/LoopVectorize/uniform-args-call-variants.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -16,17 +16,12 @@ define void @test_uniform(ptr noalias %dst, ptr readonly %src, i64 %uniform , i6
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = getelementptr double, ptr [[SRC]], i64 [[INDEX]]
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <2 x double>, ptr [[TMP0]], align 8
; CHECK-NEXT: [[TMP1:%.*]] = extractelement <2 x double> [[WIDE_LOAD]], i64 0
; CHECK-NEXT: [[TMP2:%.*]] = call double @foo(double [[TMP1]], i64 [[UNIFORM]]) #[[ATTR0:[0-9]+]]
; CHECK-NEXT: [[TMP3:%.*]] = extractelement <2 x double> [[WIDE_LOAD]], i64 1
; CHECK-NEXT: [[TMP4:%.*]] = call double @foo(double [[TMP3]], i64 [[UNIFORM]]) #[[ATTR0]]
; CHECK-NEXT: [[TMP5:%.*]] = insertelement <2 x double> poison, double [[TMP2]], i64 0
; CHECK-NEXT: [[TMP6:%.*]] = insertelement <2 x double> [[TMP5]], double [[TMP4]], i64 1
; CHECK-NEXT: [[TMP7:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: store <2 x double> [[TMP6]], ptr [[TMP7]], align 8
; CHECK-NEXT: [[TMP1:%.*]] = call <2 x double> @foo_uniform(<2 x double> [[WIDE_LOAD]], i64 [[UNIFORM]])
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDEX]]
; CHECK-NEXT: store <2 x double> [[TMP1]], ptr [[TMP2]], align 8
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 2
; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK-NEXT: [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP3]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N_VEC]], [[N]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH]]
Expand All @@ -37,7 +32,7 @@ define void @test_uniform(ptr noalias %dst, ptr readonly %src, i64 %uniform , i6
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[GEPSRC:%.*]] = getelementptr double, ptr [[SRC]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[DATA:%.*]] = load double, ptr [[GEPSRC]], align 8
; CHECK-NEXT: [[CALL:%.*]] = call double @foo(double [[DATA]], i64 [[UNIFORM]]) #[[ATTR0]]
; CHECK-NEXT: [[CALL:%.*]] = call double @foo(double [[DATA]], i64 [[UNIFORM]]) #[[ATTR0:[0-9]+]]
; CHECK-NEXT: [[GEPDST:%.*]] = getelementptr inbounds double, ptr [[DST]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store double [[CALL]], ptr [[GEPDST]], align 8
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
Expand Down

0 comments on commit 104b7c6

Please sign in to comment.