[CostModel] Handle CTLZ and CCTZ in getTypeBasedIntrinsicInstrCost

[dmgreen]

Sounds good. The costs are a little wonky for MVE (CTLZ is legal but CTTZ is custom lowered to 3 or so instructions), but they are much better then they were before!

LGTM

[samparker]

I guess these costs look more reasonable compared to their scalar counterparts, though I don't think we have cttz for MVE. But as a general query, if these are being scalarized why isn't the cost 16? It looks like it could just be the accumulation costs of the scalar parts. Or have I missed something?

[dmgreen]

It's not scalarized. Like I said it's custom lowered to a bitreverse (plus a mov) and a ctlz so 3 instructions in general.

llvm/include/llvm/CodeGen/BasicTTIImpl.h

case Intrinsic::umul_with_overflow: {
       return Cost;
     }
     case Intrinsic::ctpop:
       ISDs.push_back(ISD::CTPOP);
       // In case of legalization use TCC_Expensive. This is cheaper than a
       // library call but still not a cheap instruction.
       SingleCallCost = TargetTransformInfo::TCC_Expensive;
       break;
-    // FIXME: ctlz, cttz, ...
+    case Intrinsic::ctlz:
+      ISDs.push_back(ISD::CTLZ);
+      break;
+    case Intrinsic::cttz:
+      ISDs.push_back(ISD::CTTZ);
+      break;
     case Intrinsic::bswap:
       ISDs.push_back(ISD::BSWAP);
       break;
     case Intrinsic::bitreverse:
       ISDs.push_back(ISD::BITREVERSE);
       break;
     }
 

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

;
   %s = call float @llvm.maximum.f32(float %a, float %b)
   %v = call <16 x float> @llvm.maximum.v16f32(<16 x float> %va, <16 x float> %vb)
   ret void
 }
 
 define void @cttz(i32 %a, <16 x i32> %va) {
 ; THRU-LABEL: 'cttz'
 ; THRU-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.cttz.i32(i32 %a, i1 false)
-; THRU-NEXT:  Cost Model: Found an estimated cost of 528 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
+; THRU-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
 ; THRU-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; LATE-LABEL: 'cttz'
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.cttz.i32(i32 %a, i1 false)
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SIZE-LABEL: 'cttz'
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.cttz.i32(i32 %a, i1 false)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 528 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SIZE_LATE-LABEL: 'cttz'
 ; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.cttz.i32(i32 %a, i1 false)
-; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 528 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
+; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
 ; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %s = call i32 @llvm.cttz.i32(i32 %a, i1 false)
   %v = call <16 x i32> @llvm.cttz.v16i32(<16 x i32> %va, i1 false)
   ret void
 }
 
 define void @ctlz(i32 %a, <16 x i32> %va) {
 ; THRU-LABEL: 'ctlz'
 ; THRU-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.ctlz.i32(i32 %a, i1 true)
-; THRU-NEXT:  Cost Model: Found an estimated cost of 528 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
+; THRU-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
 ; THRU-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; LATE-LABEL: 'ctlz'
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.ctlz.i32(i32 %a, i1 true)
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
 ; LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SIZE-LABEL: 'ctlz'
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.ctlz.i32(i32 %a, i1 true)
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 528 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
 ; SIZE_LATE-LABEL: 'ctlz'
 ; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %s = call i32 @llvm.ctlz.i32(i32 %a, i1 true)
-; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 528 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
+; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
 ; SIZE_LATE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %s = call i32 @llvm.ctlz.i32(i32 %a, i1 true)
   %v = call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %va, i1 true)
   ret void
 }
 
 define void @fshl(i32 %a, i32 %b, i32 %c, <16 x i32> %va, <16 x i32> %vb, <16 x i32> %vc) {