Florian Hahn 89ae085859
[VPlan] Remove VPVectorPointer for part 0 after unrolling. (#149735)
VPVectorPointer for part 0 is just the pointer operand. Simplify it
after unrolling. This removes a large number of redundant GEPs with
index 0.

PR: https://github.com/llvm/llvm-project/pull/149735
2025-07-27 13:53:26 +01:00

93 lines
3.6 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -mtriple=riscv64 -mattr=+zve32x -passes=loop-vectorize < %s | FileCheck %s
define void @small_trip_count_min_vlen_128(ptr nocapture %a) nounwind vscale_range(4,1024) {
; CHECK-LABEL: @small_trip_count_min_vlen_128(
; CHECK-NEXT: entry:
; CHECK-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP1:%.*]], align 4
; CHECK-NEXT: [[TMP2:%.*]] = add nsw <4 x i32> [[WIDE_LOAD]], splat (i32 1)
; CHECK-NEXT: store <4 x i32> [[TMP2]], ptr [[TMP1]], align 4
; CHECK-NEXT: br label [[MIDDLE_BLOCK:%.*]]
; CHECK: middle.block:
; CHECK-NEXT: br label [[EXIT:%.*]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: br label [[LOOP1:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[IV_NEXT:%.*]], [[LOOP1]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i32 [[IV]]
; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[GEP]], align 4
; CHECK-NEXT: [[ADD:%.*]] = add nsw i32 [[V]], 1
; CHECK-NEXT: store i32 [[ADD]], ptr [[GEP]], align 4
; CHECK-NEXT: [[IV_NEXT]] = add i32 [[IV]], 1
; CHECK-NEXT: [[COND:%.*]] = icmp eq i32 [[IV]], 3
; CHECK-NEXT: br i1 [[COND]], label [[EXIT]], label [[LOOP1]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: exit:
; CHECK-NEXT: ret void
;
entry:
br label %loop
loop:
%iv = phi i32 [ %iv.next, %loop ], [ 0, %entry ]
%gep = getelementptr inbounds i32, ptr %a, i32 %iv
%v = load i32, ptr %gep, align 4
%add = add nsw i32 %v, 1
store i32 %add, ptr %gep, align 4
%iv.next = add i32 %iv, 1
%cond = icmp eq i32 %iv, 3
br i1 %cond, label %exit, label %loop
exit:
ret void
}
define void @small_trip_count_min_vlen_32(ptr nocapture %a) nounwind vscale_range(1,1024) {
; CHECK-LABEL: @small_trip_count_min_vlen_32(
; CHECK-NEXT: entry:
; CHECK-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP1:%.*]], align 4
; CHECK-NEXT: [[TMP2:%.*]] = add nsw <4 x i32> [[WIDE_LOAD]], splat (i32 1)
; CHECK-NEXT: store <4 x i32> [[TMP2]], ptr [[TMP1]], align 4
; CHECK-NEXT: br label [[MIDDLE_BLOCK:%.*]]
; CHECK: middle.block:
; CHECK-NEXT: br label [[EXIT:%.*]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: br label [[LOOP1:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[IV_NEXT:%.*]], [[LOOP1]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i32 [[IV]]
; CHECK-NEXT: [[V:%.*]] = load i32, ptr [[GEP]], align 4
; CHECK-NEXT: [[ADD:%.*]] = add nsw i32 [[V]], 1
; CHECK-NEXT: store i32 [[ADD]], ptr [[GEP]], align 4
; CHECK-NEXT: [[IV_NEXT]] = add i32 [[IV]], 1
; CHECK-NEXT: [[COND:%.*]] = icmp eq i32 [[IV]], 3
; CHECK-NEXT: br i1 [[COND]], label [[EXIT]], label [[LOOP1]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK: exit:
; CHECK-NEXT: ret void
;
entry:
br label %loop
loop:
%iv = phi i32 [ %iv.next, %loop ], [ 0, %entry ]
%gep = getelementptr inbounds i32, ptr %a, i32 %iv
%v = load i32, ptr %gep, align 4
%add = add nsw i32 %v, 1
store i32 %add, ptr %gep, align 4
%iv.next = add i32 %iv, 1
%cond = icmp eq i32 %iv, 3
br i1 %cond, label %exit, label %loop
exit:
ret void
}