[TTI][X86] getGSScalarCost - don't bother with adding cost of ICMP for each i1 mask element

These can nearly always be folded into the existing cost of the branch, and brings the throughput costs of the scalarised gather/scatter code much closer to the llvm-mca/uica estimates

Author: RKSimon

llvm/lib/Target/X86/X86TargetTransformInfo.cpp

@@ -5864,11 +5864,8 @@ InstructionCost X86TTIImpl::getGSScalarCost(unsigned Opcode,
         FixedVectorType::get(Type::getInt1Ty(SrcVTy->getContext()), VF);
     MaskUnpackCost = getScalarizationOverhead(
         MaskTy, DemandedElts, /*Insert=*/false, /*Extract=*/true, CostKind);
-    InstructionCost ScalarCompareCost = getCmpSelInstrCost(
-        Instruction::ICmp, Type::getInt1Ty(SrcVTy->getContext()), nullptr,
-        CmpInst::BAD_ICMP_PREDICATE, CostKind);
     InstructionCost BranchCost = getCFInstrCost(Instruction::Br, CostKind);
-    MaskUnpackCost += VF * (BranchCost + ScalarCompareCost);
+    MaskUnpackCost += VF * BranchCost;
   }
 
   InstructionCost AddressUnpackCost = getScalarizationOverhead(

llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll

@@ -310,7 +310,7 @@ define void @fshl(i32 %a, i32 %b, i32 %c, <16 x i32> %va, <16 x i32> %vb, <16 x
 
 define void @maskedgather(<16 x ptr> %va, <16 x i1> %vb, <16 x float> %vc) {
 ; THRU-LABEL: 'maskedgather'
-; THRU-NEXT:  Cost Model: Found an estimated cost of 77 for instruction: %v = call <16 x float> @llvm.masked.gather.v16f32.v16p0(<16 x ptr> %va, i32 1, <16 x i1> %vb, <16 x float> %vc)
+; THRU-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: %v = call <16 x float> @llvm.masked.gather.v16f32.v16p0(<16 x ptr> %va, i32 1, <16 x i1> %vb, <16 x float> %vc)
 ; THRU-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; LATE-LABEL: 'maskedgather'
@@ -331,7 +331,7 @@ define void @maskedgather(<16 x ptr> %va, <16 x i1> %vb, <16 x float> %vc) {
 
 define void @maskedscatter(<16 x float> %va, <16 x ptr> %vb, <16 x i1> %vc) {
 ; THRU-LABEL: 'maskedscatter'
-; THRU-NEXT:  Cost Model: Found an estimated cost of 77 for instruction: call void @llvm.masked.scatter.v16f32.v16p0(<16 x float> %va, <16 x ptr> %vb, i32 1, <16 x i1> %vc)
+; THRU-NEXT:  Cost Model: Found an estimated cost of 61 for instruction: call void @llvm.masked.scatter.v16f32.v16p0(<16 x float> %va, <16 x ptr> %vb, i32 1, <16 x i1> %vc)
 ; THRU-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; LATE-LABEL: 'maskedscatter'

llvm/test/Analysis/CostModel/X86/masked-gather-i32-with-i8-index.ll

@@ -49,8 +49,8 @@ define void @test() {
 ;
 ; AVX512-LABEL: 'test'
 ; AVX512:  LV: Found an estimated cost of 1 for VF 1 For instruction: %valB.loaded = load i32, ptr %inB, align 4
-; AVX512:  LV: Found an estimated cost of 10 for VF 2 For instruction: %valB.loaded = load i32, ptr %inB, align 4
-; AVX512:  LV: Found an estimated cost of 21 for VF 4 For instruction: %valB.loaded = load i32, ptr %inB, align 4
+; AVX512:  LV: Found an estimated cost of 8 for VF 2 For instruction: %valB.loaded = load i32, ptr %inB, align 4
+; AVX512:  LV: Found an estimated cost of 17 for VF 4 For instruction: %valB.loaded = load i32, ptr %inB, align 4
 ; AVX512:  LV: Found an estimated cost of 10 for VF 8 For instruction: %valB.loaded = load i32, ptr %inB, align 4
 ; AVX512:  LV: Found an estimated cost of 18 for VF 16 For instruction: %valB.loaded = load i32, ptr %inB, align 4
 ; AVX512:  LV: Found an estimated cost of 36 for VF 32 For instruction: %valB.loaded = load i32, ptr %inB, align 4

llvm/test/Analysis/CostModel/X86/masked-gather-i64-with-i8-index.ll

@@ -49,8 +49,8 @@ define void @test() {
 ;
 ; AVX512-LABEL: 'test'
 ; AVX512:  LV: Found an estimated cost of 1 for VF 1 For instruction: %valB.loaded = load i64, ptr %inB, align 8
-; AVX512:  LV: Found an estimated cost of 10 for VF 2 For instruction: %valB.loaded = load i64, ptr %inB, align 8
-; AVX512:  LV: Found an estimated cost of 22 for VF 4 For instruction: %valB.loaded = load i64, ptr %inB, align 8
+; AVX512:  LV: Found an estimated cost of 8 for VF 2 For instruction: %valB.loaded = load i64, ptr %inB, align 8
+; AVX512:  LV: Found an estimated cost of 18 for VF 4 For instruction: %valB.loaded = load i64, ptr %inB, align 8
 ; AVX512:  LV: Found an estimated cost of 10 for VF 8 For instruction: %valB.loaded = load i64, ptr %inB, align 8
 ; AVX512:  LV: Found an estimated cost of 20 for VF 16 For instruction: %valB.loaded = load i64, ptr %inB, align 8
 ; AVX512:  LV: Found an estimated cost of 40 for VF 32 For instruction: %valB.loaded = load i64, ptr %inB, align 8