[RISCV][CostModel] Make VMV_S_* and VMV_*_S cost independent of LMUL

[arcbbb]

Following #77963, instructions like VMV_S_X/VMV_X_S/VFMV_S_F/VFMV_F_S handle single element, so the cost don't scale with LMUL.

[llvmbot]

@llvm/pr-subscribers-backend-risc-v

@llvm/pr-subscribers-llvm-analysis

Author: Shih-Po Hung (arcbbb)

Changes

Following #77963, instructions like VMV_S_X/VMV_X_S/VFMV_S_F/VFMV_F_S handle single element, so the cost don't scale with LMUL.

---

Full diff: https://github.com/llvm/llvm-project/pull/78739.diff

3 Files Affected:

• (modified) llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp (+9-4)
• (modified) llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll (+2-2)
• (modified) llvm/test/Analysis/CostModel/RISCV/shuffle-broadcast.ll (+6-6)

diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index 4ea3a519308995..71fd6e96316140 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -84,6 +84,12 @@ RISCVTTIImpl::getRISCVInstructionCost(ArrayRef<unsigned> OpCodes, MVT VT,
       Cost += VL;
       break;
     }
+    case RISCV::VMV_X_S:
+    case RISCV::VFMV_F_S:
+    case RISCV::VMV_S_X:
+    case RISCV::VFMV_S_F:
+      Cost += 1;
+      break;
     default:
       Cost += LMULCost;
     }
@@ -446,7 +452,7 @@ InstructionCost RISCVTTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
     // should be a very small constant for the constant pool load.  As such,
     // we may bias towards large selects slightly more than truely warranted.
     return LT.first *
-           (2 + getRISCVInstructionCost({RISCV::VMERGE_VVM},
+           (1 + getRISCVInstructionCost({RISCV::VMV_S_X, RISCV::VMERGE_VVM},
                                         LT.second, CostKind));
   }
   case TTI::SK_Broadcast: {
@@ -475,10 +481,9 @@ InstructionCost RISCVTTIImpl::getShuffleCost(TTI::ShuffleKind Kind,
 
       return LT.first *
              (TLI->getLMULCost(LT.second) + // FIXME: this should be 1 for andi
-              TLI->getLMULCost(
-                  LT.second) + // FIXME: vmv.x.s is the same as extractelement
               getRISCVInstructionCost({RISCV::VMV_V_I, RISCV::VMERGE_VIM,
-                                       RISCV::VMV_V_X, RISCV::VMSNE_VI},
+                                       RISCV::VMV_X_S, RISCV::VMV_V_X,
+                                       RISCV::VMSNE_VI},
                                       LT.second, CostKind));
     }
 
diff --git a/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll b/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
index bd9f6af89a5cd9..4f3c7e2f90c655 100644
--- a/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/rvv-shuffle.ll
@@ -14,7 +14,7 @@ define void  @vector_broadcast() {
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %5 = shufflevector <vscale x 4 x i32> undef, <vscale x 4 x i32> undef, <vscale x 4 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %6 = shufflevector <vscale x 1 x i64> undef, <vscale x 1 x i64> undef, <vscale x 1 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %7 = shufflevector <vscale x 2 x i64> undef, <vscale x 2 x i64> undef, <vscale x 2 x i32> zeroinitializer
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %8 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %8 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %9 = shufflevector <vscale x 8 x i1> undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %10 = shufflevector <vscale x 4 x i1> undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %11 = shufflevector <vscale x 2 x i1> undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer
@@ -29,7 +29,7 @@ define void  @vector_broadcast() {
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %5 = shufflevector <vscale x 4 x i32> undef, <vscale x 4 x i32> undef, <vscale x 4 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %6 = shufflevector <vscale x 1 x i64> undef, <vscale x 1 x i64> undef, <vscale x 1 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %7 = shufflevector <vscale x 2 x i64> undef, <vscale x 2 x i64> undef, <vscale x 2 x i32> zeroinitializer
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %8 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %8 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %9 = shufflevector <vscale x 8 x i1> undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %10 = shufflevector <vscale x 4 x i1> undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %11 = shufflevector <vscale x 2 x i1> undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer
diff --git a/llvm/test/Analysis/CostModel/RISCV/shuffle-broadcast.ll b/llvm/test/Analysis/CostModel/RISCV/shuffle-broadcast.ll
index 432b90d9305af3..fc4a6b17d3f826 100644
--- a/llvm/test/Analysis/CostModel/RISCV/shuffle-broadcast.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/shuffle-broadcast.ll
@@ -45,9 +45,9 @@ define void  @broadcast_scalable() #0{
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %38 = shufflevector <vscale x 2 x i1> undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %39 = shufflevector <vscale x 4 x i1> undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %40 = shufflevector <vscale x 8 x i1> undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> zeroinitializer
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %41 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %42 = shufflevector <vscale x 32 x i1> undef, <vscale x 32 x i1> undef, <vscale x 32 x i32> zeroinitializer
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %43 = shufflevector <vscale x 64 x i1> undef, <vscale x 64 x i1> undef, <vscale x 64 x i32> zeroinitializer
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 11 for instruction: %41 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 21 for instruction: %42 = shufflevector <vscale x 32 x i1> undef, <vscale x 32 x i1> undef, <vscale x 32 x i32> zeroinitializer
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 41 for instruction: %43 = shufflevector <vscale x 64 x i1> undef, <vscale x 64 x i1> undef, <vscale x 64 x i32> zeroinitializer
 ; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;
 ; SIZE-LABEL: 'broadcast_scalable'
@@ -92,9 +92,9 @@ define void  @broadcast_scalable() #0{
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %38 = shufflevector <vscale x 2 x i1> undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %39 = shufflevector <vscale x 4 x i1> undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %40 = shufflevector <vscale x 8 x i1> undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> zeroinitializer
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %41 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %42 = shufflevector <vscale x 32 x i1> undef, <vscale x 32 x i1> undef, <vscale x 32 x i32> zeroinitializer
-; SIZE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %43 = shufflevector <vscale x 64 x i1> undef, <vscale x 64 x i1> undef, <vscale x 64 x i32> zeroinitializer
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %41 = shufflevector <vscale x 16 x i1> undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> zeroinitializer
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %42 = shufflevector <vscale x 32 x i1> undef, <vscale x 32 x i1> undef, <vscale x 32 x i32> zeroinitializer
+; SIZE-NEXT:  Cost Model: Found an estimated cost of 13 for instruction: %43 = shufflevector <vscale x 64 x i1> undef, <vscale x 64 x i1> undef, <vscale x 64 x i32> zeroinitializer
 ; SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret void
 ;
   %zero = shufflevector <vscale x 1 x half> undef, <vscale x 1 x half> undef, <vscale x 1 x i32> zeroinitializer

[wangpc-pp]

LGTM.

[preames]

LGTM w/comment addressed.

[preames]

VFMV_F_S and VFMV_S_F are not used in this patch, please remove.

[preames]

Off topic for the actual change, but I think we can improve the code sequence above.

One idea would be to use a scalar multiply to do the broadcast since we know the value will either be 0 or (int16)-1.
Another would be to use a vrgather.vi in place of the extract/splat sequence.
Another would be to extract the full16 bit value, do the extract and splat in scalar, and insert a full 16 bit value back.

Not sure if you care about this case or not, I don't much.

[arcbbb]

Understood! Let's consider revisiting it when it becomes more relevant!

[preames]

Hm, one thing I'm noticing. I think the example code sequence above is only for short fixed vectors. For that to work on scalable vectors, I think we need a slide in there somewhere. Or, said alternatively, that remaining LMUL scaling is probably modeling the vslidedown more than the andi for scalable vectors and we need to adjust our example and comments slightly.

(Not directly relevant to your change, more the existing code structure.)