[llvm] [LV] Bundle (partial) reductions with a mul of a constant (PR #162503)

[Sander de Smalen via llvm-commits]

================
@@ -2800,6 +2800,91 @@ exit:
   ret i64 %r.0.lcssa
 }
 
+define i32 @reduction_expression_ext_mulacc_livein(ptr %a, ptr %b, i16 %c) {
+; CHECK-LABEL: define i32 @reduction_expression_ext_mulacc_livein(
+; CHECK-SAME: ptr [[A:%.*]], ptr [[B:%.*]], i16 [[C:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
+; CHECK:       [[VECTOR_PH]]:
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i16> poison, i16 [[C]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i16> [[BROADCAST_SPLATINSERT]], <4 x i16> poison, <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]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[TMP5:%.*]], %[[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i8, ptr [[A]], i64 [[INDEX]]
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, ptr [[TMP0]], align 1
+; CHECK-NEXT:    [[TMP1:%.*]] = zext <4 x i8> [[WIDE_LOAD]] to <4 x i16>
+; CHECK-NEXT:    [[TMP2:%.*]] = mul <4 x i16> [[BROADCAST_SPLAT]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
+; CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP3]])
+; CHECK-NEXT:    [[TMP5]] = add i32 [[VEC_PHI]], [[TMP4]]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1024
+; CHECK-NEXT:    br i1 [[TMP6]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP32:![0-9]+]]
+; CHECK:       [[MIDDLE_BLOCK]]:
+; CHECK-NEXT:    br label %[[FOR_EXIT:.*]]
+; CHECK:       [[FOR_EXIT]]:
+; CHECK-NEXT:    ret i32 [[TMP5]]
+;
+; CHECK-INTERLEAVED-LABEL: define i32 @reduction_expression_ext_mulacc_livein(
+; CHECK-INTERLEAVED-SAME: ptr [[A:%.*]], ptr [[B:%.*]], i16 [[C:%.*]]) {
+; CHECK-INTERLEAVED-NEXT:  [[ENTRY:.*:]]
+; CHECK-INTERLEAVED-NEXT:    br label %[[VECTOR_PH:.*]]
+; CHECK-INTERLEAVED:       [[VECTOR_PH]]:
+; CHECK-INTERLEAVED-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i16> poison, i16 [[C]], i64 0
+; CHECK-INTERLEAVED-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i16> [[BROADCAST_SPLATINSERT]], <4 x i16> poison, <4 x i32> zeroinitializer
+; CHECK-INTERLEAVED-NEXT:    br label %[[VECTOR_BODY:.*]]
+; CHECK-INTERLEAVED:       [[VECTOR_BODY]]:
+; CHECK-INTERLEAVED-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
+; CHECK-INTERLEAVED-NEXT:    [[VEC_PHI:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[TMP8:%.*]], %[[VECTOR_BODY]] ]
+; CHECK-INTERLEAVED-NEXT:    [[VEC_PHI1:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[TMP11:%.*]], %[[VECTOR_BODY]] ]
+; CHECK-INTERLEAVED-NEXT:    [[TMP0:%.*]] = getelementptr i8, ptr [[A]], i64 [[INDEX]]
+; CHECK-INTERLEAVED-NEXT:    [[TMP1:%.*]] = getelementptr i8, ptr [[TMP0]], i32 4
+; CHECK-INTERLEAVED-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i8>, ptr [[TMP0]], align 1
+; CHECK-INTERLEAVED-NEXT:    [[WIDE_LOAD2:%.*]] = load <4 x i8>, ptr [[TMP1]], align 1
+; CHECK-INTERLEAVED-NEXT:    [[TMP2:%.*]] = zext <4 x i8> [[WIDE_LOAD]] to <4 x i16>
+; CHECK-INTERLEAVED-NEXT:    [[TMP3:%.*]] = zext <4 x i8> [[WIDE_LOAD2]] to <4 x i16>
+; CHECK-INTERLEAVED-NEXT:    [[TMP4:%.*]] = mul <4 x i16> [[BROADCAST_SPLAT]], [[TMP2]]
+; CHECK-INTERLEAVED-NEXT:    [[TMP5:%.*]] = mul <4 x i16> [[BROADCAST_SPLAT]], [[TMP3]]
+; CHECK-INTERLEAVED-NEXT:    [[TMP6:%.*]] = zext <4 x i16> [[TMP4]] to <4 x i32>
+; CHECK-INTERLEAVED-NEXT:    [[TMP7:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP6]])
+; CHECK-INTERLEAVED-NEXT:    [[TMP8]] = add i32 [[VEC_PHI]], [[TMP7]]
+; CHECK-INTERLEAVED-NEXT:    [[TMP9:%.*]] = zext <4 x i16> [[TMP5]] to <4 x i32>
+; CHECK-INTERLEAVED-NEXT:    [[TMP10:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[TMP9]])
+; CHECK-INTERLEAVED-NEXT:    [[TMP11]] = add i32 [[VEC_PHI1]], [[TMP10]]
+; CHECK-INTERLEAVED-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
+; CHECK-INTERLEAVED-NEXT:    [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1024
+; CHECK-INTERLEAVED-NEXT:    br i1 [[TMP12]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP32:![0-9]+]]
+; CHECK-INTERLEAVED:       [[MIDDLE_BLOCK]]:
+; CHECK-INTERLEAVED-NEXT:    [[BIN_RDX:%.*]] = add i32 [[TMP11]], [[TMP8]]
+; CHECK-INTERLEAVED-NEXT:    br label %[[FOR_EXIT:.*]]
+; CHECK-INTERLEAVED:       [[FOR_EXIT]]:
+; CHECK-INTERLEAVED-NEXT:    ret i32 [[BIN_RDX]]
+;
+entry:
+  br label %for.body
+
+for.body:                                         ; preds = %for.body, %entry
+  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
+  %accum = phi i32 [ 0, %entry ], [ %add, %for.body ]
+  %gep.a = getelementptr i8, ptr %a, i64 %iv
+  %load.a = load i8, ptr %gep.a, align 1
+  %ext.a = zext i8 %load.a to i16
+  %gep.b = getelementptr i8, ptr %b, i64 %iv
+  %load.b = load i8, ptr %gep.b, align 1
+  %ext.b = zext i8 %load.b to i16
----------------
sdesmalen-arm wrote:

these are unused (and so is argument `ptr %b`)

https://github.com/llvm/llvm-project/pull/162503