[llvm] [VPlan] Mark histogram increment as only having its first lane used. (PR #226138)

[Florian Hahn via llvm-commits]

https://github.com/fhahn created https://github.com/llvm/llvm-project/pull/226138

The second operand (increment ) of llvm.experimental.vector.histogram.* is a scalar. Handle it in usesFirstLaneOnly.

>From b62a51e1215a866b79f1514b5e28a1d239d45188 Mon Sep 17 00:00:00 2001
From: Florian Hahn <flo at fhahn.com>
Date: Thu, 24 Sep 2026 10:57:28 +0100
Subject: [PATCH] [VPlan] Mark histogram increment as only having its first
 lane used.

The second operand (increment ) of llvm.experimental.vector.histogram.* is a
scalar. Handle it in usesFirstLaneOnly.
---
 llvm/lib/Transforms/Vectorize/VPlan.h                      | 7 +++++++
 llvm/test/Transforms/LoopVectorize/AArch64/sve2-histcnt.ll | 5 +----
 2 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h
index 8f22a362e6fe6b..30af5f576444ea 100644
--- a/llvm/lib/Transforms/Vectorize/VPlan.h
+++ b/llvm/lib/Transforms/Vectorize/VPlan.h
@@ -2206,6 +2206,13 @@ class VPHistogramRecipe : public VPRecipeBase, public VPIRMetadata {
     return getNumOperands() == 3 ? getOperand(2) : nullptr;
   }
 
+  /// Returns true if the recipe only uses the first lane of operand \p Op.
+  bool usesFirstLaneOnly(const VPValue *Op) const override {
+    assert(is_contained(operands(), Op) &&
+           "Op must be an operand of the recipe");
+    return Op == getOperand(1);
+  }
+
 protected:
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
   /// Print the recipe
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/sve2-histcnt.ll b/llvm/test/Transforms/LoopVectorize/AArch64/sve2-histcnt.ll
index ca92a8e372ea11..f929e3b5d80a46 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/sve2-histcnt.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/sve2-histcnt.ll
@@ -85,8 +85,6 @@ define void @simple_histogram_inc_param(ptr noalias %buckets, ptr readonly %indi
 ; CHECK:       vector.ph:
 ; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP1]]
 ; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
-; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[INCVAL]], i64 0
-; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[BROADCAST_SPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
@@ -94,8 +92,7 @@ define void @simple_histogram_inc_param(ptr noalias %buckets, ptr readonly %indi
 ; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, ptr [[TMP8]], align 4
 ; CHECK-NEXT:    [[TMP9:%.*]] = zext <vscale x 4 x i32> [[WIDE_LOAD]] to <vscale x 4 x i64>
 ; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, ptr [[BUCKETS]], <vscale x 4 x i64> [[TMP9]]
-; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <vscale x 4 x i32> [[BROADCAST_SPLAT]], i64 0
-; CHECK-NEXT:    call void @llvm.experimental.vector.histogram.add.nxv4p0.i32(<vscale x 4 x ptr> [[TMP6]], i32 [[TMP7]], <vscale x 4 x i1> splat (i1 true))
+; CHECK-NEXT:    call void @llvm.experimental.vector.histogram.add.nxv4p0.i32(<vscale x 4 x ptr> [[TMP6]], i32 [[INCVAL]], <vscale x 4 x i1> splat (i1 true))
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP1]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]