Florian Hahn 50b9ca4dda
[VPlan] Simplify Plan's entry in removeBranchOnConst. (#154510)
After https://github.com/llvm/llvm-project/pull/153643, there may be a
BranchOnCond with constant condition in the entry block.

Simplify those in removeBranchOnConst. This removes a number of
redundant conditional branch from entry blocks.

In some cases, it may also make the original scalar loop unreachable,
because we know it will never execute. In that case, we need to remove
the loop from LoopInfo, because all unreachable blocks may dominate each
other, making LoopInfo invalid. In those cases, we can also completely
remove the loop, for which I'll share a follow-up patch.

Depends on https://github.com/llvm/llvm-project/pull/153643.

PR: https://github.com/llvm/llvm-project/pull/154510
2025-09-18 19:25:05 +01:00

73 lines
3.5 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -passes=loop-vectorize -force-vector-width=4 -force-vector-interleave=1 < %s | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128-ni:1"
define void @foo(ptr %ptr, ptr %ptr.2) {
; CHECK-LABEL: @foo(
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[VECTOR_MEMCHECK:%.*]]
; CHECK: vector.memcheck:
; CHECK-NEXT: [[UGLYGEP:%.*]] = getelementptr i8, ptr [[PTR_2:%.*]], i64 4
; CHECK-NEXT: [[UGLYGEP1:%.*]] = getelementptr i8, ptr [[PTR:%.*]], i64 640
; CHECK-NEXT: [[BOUND0:%.*]] = icmp ult ptr [[PTR_2]], [[UGLYGEP1]]
; CHECK-NEXT: [[BOUND1:%.*]] = icmp ult ptr [[PTR]], [[UGLYGEP]]
; CHECK-NEXT: [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]]
; CHECK-NEXT: br i1 [[FOUND_CONFLICT]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VEC_IND:%.*]] = phi <4 x i64> [ <i64 2, i64 3, i64 4, i64 5>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = add i64 2, [[INDEX]]
; CHECK-NEXT: [[TMP0:%.*]] = trunc i64 [[OFFSET_IDX]] to i32
; CHECK-NEXT: [[TMP1:%.*]] = add i32 [[TMP0]], 0
; CHECK-NEXT: [[TMP2:%.*]] = add i32 [[TMP0]], 1
; CHECK-NEXT: [[TMP3:%.*]] = add i32 [[TMP0]], 2
; CHECK-NEXT: [[TMP4:%.*]] = add i32 [[TMP0]], 3
; CHECK-NEXT: store i32 [[TMP4]], ptr [[PTR_2]], align 4, !alias.scope [[META0:![0-9]+]], !noalias [[META3:![0-9]+]]
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i64, ptr [[PTR]], i64 [[INDEX]]
; CHECK-NEXT: store <4 x i64> [[VEC_IND]], ptr [[TMP6]], align 8, !alias.scope [[META3]]
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <4 x i64> [[VEC_IND]], splat (i64 4)
; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], 80
; CHECK-NEXT: br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: br label [[EXIT:%.*]]
; CHECK: scalar.ph:
; CHECK-NEXT: br label [[LOOP:%.*]]
; CHECK: loop:
; CHECK-NEXT: [[CAN_IV:%.*]] = phi i64 [ 0, [[SCALAR_PH]] ], [ [[CAN_IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[TMP9:%.*]] = phi i64 [ 2, [[SCALAR_PH]] ], [ [[TMP12:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[TMP10:%.*]] = and i64 [[TMP9]], 4294967295
; CHECK-NEXT: [[TMP11:%.*]] = trunc i64 [[TMP9]] to i32
; CHECK-NEXT: store i32 [[TMP11]], ptr [[PTR_2]], align 4
; CHECK-NEXT: [[GEP_PTR:%.*]] = getelementptr inbounds i64, ptr [[PTR]], i64 [[CAN_IV]]
; CHECK-NEXT: store i64 [[TMP9]], ptr [[GEP_PTR]], align 8
; CHECK-NEXT: [[TMP12]] = add nuw nsw i64 [[TMP10]], 1
; CHECK-NEXT: [[TMP13:%.*]] = icmp sgt i32 [[TMP11]], 80
; CHECK-NEXT: [[CAN_IV_NEXT]] = add nuw nsw i64 [[CAN_IV]], 1
; CHECK-NEXT: br i1 [[TMP13]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP8:![0-9]+]]
; CHECK: exit:
; CHECK-NEXT: ret void
;
entry:
br label %loop
loop:
%can.iv = phi i64 [ 0, %entry ], [ %can.iv.next, %loop ]
%0 = phi i64 [ 2, %entry ], [ %3, %loop ]
%1 = and i64 %0, 4294967295
%2 = trunc i64 %0 to i32
store i32 %2, ptr %ptr.2
%gep.ptr = getelementptr inbounds i64, ptr %ptr, i64 %can.iv
store i64 %0, ptr %gep.ptr
%3 = add nuw nsw i64 %1, 1
%4 = icmp sgt i32 %2, 80
%can.iv.next = add nuw nsw i64 %can.iv, 1
br i1 %4, label %exit, label %loop
exit:
ret void
}