86813aa786
Epilogue vectorization currently relies on the resume phi for the canonical induction being always available, which is why VPPhi are considered to have side-effects, to prevent their removal. This patch adds a new ResumeForEpilogue opcode to mark the resume phi as used for epilogue vectorization. This allows treating VPPhis in general as not having side-effects, enabling removal of unused VPPhis.
210 lines
11 KiB
LLVM
210 lines
11 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: opt < %s -force-vector-width=2 -force-vector-interleave=2 -passes=loop-vectorize -S | FileCheck %s
|
|
|
|
target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
|
|
|
|
;
|
|
; This test ensures that we don't generate trivially dead instructions prior to
|
|
; instruction simplification. We don't need to generate instructions
|
|
; corresponding to the original induction variable update or branch condition,
|
|
; since we rewrite the loop structure.
|
|
;
|
|
define i64 @dead_instructions_01(ptr %a, i64 %n) {
|
|
; CHECK-LABEL: define i64 @dead_instructions_01(
|
|
; CHECK-SAME: ptr [[A:%.*]], i64 [[N:%.*]]) {
|
|
; CHECK-NEXT: [[ENTRY:.*]]:
|
|
; CHECK-NEXT: [[SMAX:%.*]] = call i64 @llvm.smax.i64(i64 [[N]], i64 1)
|
|
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[SMAX]], 4
|
|
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
|
|
; CHECK: [[VECTOR_PH]]:
|
|
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[SMAX]], 4
|
|
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[SMAX]], [[N_MOD_VF]]
|
|
; 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 <2 x i64> [ zeroinitializer, %[[VECTOR_PH]] ], [ [[TMP6:%.*]], %[[VECTOR_BODY]] ]
|
|
; CHECK-NEXT: [[VEC_PHI1:%.*]] = phi <2 x i64> [ zeroinitializer, %[[VECTOR_PH]] ], [ [[TMP7:%.*]], %[[VECTOR_BODY]] ]
|
|
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[INDEX]]
|
|
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i64, ptr [[TMP2]], i32 2
|
|
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <2 x i64>, ptr [[TMP2]], align 8
|
|
; CHECK-NEXT: [[WIDE_LOAD2:%.*]] = load <2 x i64>, ptr [[TMP5]], align 8
|
|
; CHECK-NEXT: [[TMP6]] = add <2 x i64> [[WIDE_LOAD]], [[VEC_PHI]]
|
|
; CHECK-NEXT: [[TMP7]] = add <2 x i64> [[WIDE_LOAD2]], [[VEC_PHI1]]
|
|
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
|
|
; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
|
|
; CHECK-NEXT: br i1 [[TMP8]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
|
|
; CHECK: [[MIDDLE_BLOCK]]:
|
|
; CHECK-NEXT: [[BIN_RDX:%.*]] = add <2 x i64> [[TMP7]], [[TMP6]]
|
|
; CHECK-NEXT: [[TMP9:%.*]] = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> [[BIN_RDX]])
|
|
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[SMAX]], [[N_VEC]]
|
|
; CHECK-NEXT: br i1 [[CMP_N]], label %[[FOR_END:.*]], label %[[SCALAR_PH]]
|
|
; CHECK: [[SCALAR_PH]]:
|
|
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ]
|
|
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i64 [ [[TMP9]], %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ]
|
|
; CHECK-NEXT: br label %[[FOR_BODY:.*]]
|
|
; CHECK: [[FOR_BODY]]:
|
|
; CHECK-NEXT: [[I:%.*]] = phi i64 [ [[I_NEXT:%.*]], %[[FOR_BODY]] ], [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ]
|
|
; CHECK-NEXT: [[R:%.*]] = phi i64 [ [[TMP6:%.*]], %[[FOR_BODY]] ], [ [[BC_MERGE_RDX]], %[[SCALAR_PH]] ]
|
|
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[I]]
|
|
; CHECK-NEXT: [[TMP5:%.*]] = load i64, ptr [[TMP2]], align 8
|
|
; CHECK-NEXT: [[TMP6]] = add i64 [[TMP5]], [[R]]
|
|
; CHECK-NEXT: [[I_NEXT]] = add nuw nsw i64 [[I]], 1
|
|
; CHECK-NEXT: [[COND:%.*]] = icmp slt i64 [[I_NEXT]], [[N]]
|
|
; CHECK-NEXT: br i1 [[COND]], label %[[FOR_BODY]], label %[[FOR_END]], !llvm.loop [[LOOP3:![0-9]+]]
|
|
; CHECK: [[FOR_END]]:
|
|
; CHECK-NEXT: [[TMP7:%.*]] = phi i64 [ [[TMP6]], %[[FOR_BODY]] ], [ [[TMP9]], %[[MIDDLE_BLOCK]] ]
|
|
; CHECK-NEXT: ret i64 [[TMP7]]
|
|
;
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%i = phi i64 [ %i.next, %for.body ], [ 0, %entry ]
|
|
%r = phi i64 [ %tmp2, %for.body ], [ 0, %entry ]
|
|
%tmp0 = getelementptr inbounds i64, ptr %a, i64 %i
|
|
%tmp1 = load i64, ptr %tmp0, align 8
|
|
%tmp2 = add i64 %tmp1, %r
|
|
%i.next = add nuw nsw i64 %i, 1
|
|
%cond = icmp slt i64 %i.next, %n
|
|
br i1 %cond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
%tmp3 = phi i64 [ %tmp2, %for.body ]
|
|
ret i64 %tmp3
|
|
}
|
|
|
|
|
|
;
|
|
; This test ensures that the primary induction is not considered dead when
|
|
; acting as the 'add' of another induction, and otherwise feeding only its own
|
|
; 'add' (recognized earlier as 'dead'), when the tail of the loop is folded by
|
|
; masking. Such masking uses the primary induction.
|
|
;
|
|
define void @pr47390(ptr %a) {
|
|
; CHECK-LABEL: define void @pr47390(
|
|
; CHECK-SAME: ptr [[A:%.*]]) {
|
|
; CHECK-NEXT: [[ENTRY:.*:]]
|
|
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_PH:.*]]
|
|
; CHECK: [[VECTOR_PH]]:
|
|
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
|
|
; CHECK: [[VECTOR_BODY]]:
|
|
; CHECK-NEXT: [[INDEX:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
|
|
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
|
|
; CHECK-NEXT: [[TMP0:%.*]] = icmp eq i32 [[INDEX_NEXT]], 8
|
|
; CHECK-NEXT: br i1 [[TMP0]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
|
|
; CHECK: [[MIDDLE_BLOCK]]:
|
|
; CHECK-NEXT: br label %[[EXIT:.*]]
|
|
; CHECK: [[SCALAR_PH]]:
|
|
; CHECK-NEXT: br label %[[LOOP:.*]]
|
|
; CHECK: [[EXIT]]:
|
|
; CHECK-NEXT: ret void
|
|
; CHECK: [[LOOP]]:
|
|
; CHECK-NEXT: [[PRIMARY:%.*]] = phi i32 [ 0, %[[SCALAR_PH]] ], [ [[PRIMARY_ADD:%.*]], %[[LOOP]] ]
|
|
; CHECK-NEXT: [[USE_PRIMARY:%.*]] = phi i32 [ -1, %[[SCALAR_PH]] ], [ [[PRIMARY]], %[[LOOP]] ]
|
|
; CHECK-NEXT: [[SECONDARY:%.*]] = phi i32 [ 1, %[[SCALAR_PH]] ], [ [[SECONDARY_ADD:%.*]], %[[LOOP]] ]
|
|
; CHECK-NEXT: [[PRIMARY_ADD]] = add i32 [[PRIMARY]], 1
|
|
; CHECK-NEXT: [[SECONDARY_ADD]] = add i32 [[SECONDARY]], 1
|
|
; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[A]], i32 [[SECONDARY]]
|
|
; CHECK-NEXT: [[LOAD:%.*]] = load i32, ptr [[GEP]], align 8
|
|
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[SECONDARY]], 5
|
|
; CHECK-NEXT: br i1 [[CMP]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP5:![0-9]+]]
|
|
;
|
|
entry:
|
|
br label %loop
|
|
|
|
exit:
|
|
ret void
|
|
|
|
loop:
|
|
%primary = phi i32 [ 0, %entry ], [ %primary_add, %loop ]
|
|
%use_primary = phi i32 [ -1, %entry ], [ %primary, %loop ]
|
|
%secondary = phi i32 [ 1, %entry ], [ %secondary_add, %loop ]
|
|
%primary_add = add i32 %primary, 1
|
|
%secondary_add = add i32 %secondary, 1
|
|
%gep = getelementptr inbounds i32, ptr %a, i32 %secondary
|
|
%load = load i32, ptr %gep, align 8
|
|
%cmp = icmp eq i32 %secondary, 5
|
|
br i1 %cmp, label %exit, label %loop
|
|
}
|
|
|
|
; Test with a dead load and dead vector poiner.
|
|
define void @dead_load_and_vector_pointer(ptr %a, ptr %b) {
|
|
; CHECK-LABEL: define void @dead_load_and_vector_pointer(
|
|
; CHECK-SAME: ptr [[A:%.*]], ptr [[B:%.*]]) {
|
|
; CHECK-NEXT: [[ENTRY:.*]]:
|
|
; CHECK-NEXT: br i1 false, label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]]
|
|
; CHECK: [[VECTOR_MEMCHECK]]:
|
|
; CHECK-NEXT: [[SCEVGEP:%.*]] = getelementptr i8, ptr [[A]], i64 516
|
|
; CHECK-NEXT: [[SCEVGEP1:%.*]] = getelementptr i8, ptr [[B]], i64 516
|
|
; CHECK-NEXT: [[BOUND0:%.*]] = icmp ult ptr [[A]], [[SCEVGEP1]]
|
|
; CHECK-NEXT: [[BOUND1:%.*]] = icmp ult ptr [[B]], [[SCEVGEP]]
|
|
; 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 i32 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
|
|
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[A]], i32 [[INDEX]]
|
|
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i32, ptr [[TMP2]], i32 2
|
|
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <2 x i32>, ptr [[TMP2]], align 8, !alias.scope [[META6:![0-9]+]], !noalias [[META9:![0-9]+]]
|
|
; CHECK-NEXT: [[WIDE_LOAD2:%.*]] = load <2 x i32>, ptr [[TMP5]], align 8, !alias.scope [[META6]], !noalias [[META9]]
|
|
; CHECK-NEXT: [[TMP6:%.*]] = add <2 x i32> [[WIDE_LOAD]], splat (i32 1)
|
|
; CHECK-NEXT: [[TMP7:%.*]] = add <2 x i32> [[WIDE_LOAD2]], splat (i32 1)
|
|
; CHECK-NEXT: store <2 x i32> [[TMP6]], ptr [[TMP2]], align 4, !alias.scope [[META6]], !noalias [[META9]]
|
|
; CHECK-NEXT: store <2 x i32> [[TMP7]], ptr [[TMP5]], align 4, !alias.scope [[META6]], !noalias [[META9]]
|
|
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
|
|
; CHECK-NEXT: [[TMP12:%.*]] = icmp eq i32 [[INDEX_NEXT]], 128
|
|
; CHECK-NEXT: br i1 [[TMP12]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
|
|
; CHECK: [[MIDDLE_BLOCK]]:
|
|
; CHECK-NEXT: br label %[[SCALAR_PH]]
|
|
; CHECK: [[SCALAR_PH]]:
|
|
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ 128, %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ], [ 0, %[[VECTOR_MEMCHECK]] ]
|
|
; CHECK-NEXT: br label %[[LOOP:.*]]
|
|
; CHECK: [[EXIT:.*]]:
|
|
; CHECK-NEXT: ret void
|
|
; CHECK: [[LOOP]]:
|
|
; CHECK-NEXT: [[PRIMARY:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[PRIMARY_ADD:%.*]], %[[LOOP]] ]
|
|
; CHECK-NEXT: [[PRIMARY_ADD]] = add i32 [[PRIMARY]], 1
|
|
; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[A]], i32 [[PRIMARY]]
|
|
; CHECK-NEXT: [[LOAD:%.*]] = load i32, ptr [[GEP]], align 8
|
|
; CHECK-NEXT: [[ADD:%.*]] = add i32 [[LOAD]], 1
|
|
; CHECK-NEXT: store i32 [[ADD]], ptr [[GEP]], align 4
|
|
; CHECK-NEXT: [[GEP_B:%.*]] = getelementptr inbounds i32, ptr [[B]], i32 [[PRIMARY]]
|
|
; CHECK-NEXT: [[LOAD2:%.*]] = load i32, ptr [[GEP_B]], align 4
|
|
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[PRIMARY]], 128
|
|
; CHECK-NEXT: br i1 [[CMP]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP12:![0-9]+]]
|
|
;
|
|
entry:
|
|
br label %loop
|
|
|
|
exit:
|
|
ret void
|
|
|
|
loop:
|
|
%primary = phi i32 [ 0, %entry ], [ %primary_add, %loop ]
|
|
%primary_add = add i32 %primary, 1
|
|
%gep = getelementptr inbounds i32, ptr %a, i32 %primary
|
|
%load = load i32, ptr %gep, align 8
|
|
%add = add i32 %load, 1
|
|
store i32 %add, ptr %gep
|
|
%gep.b = getelementptr inbounds i32, ptr %b, i32 %primary
|
|
%load2 = load i32, ptr %gep.b
|
|
%cmp = icmp eq i32 %primary, 128
|
|
br i1 %cmp, label %exit, label %loop
|
|
}
|
|
;.
|
|
; CHECK: [[LOOP0]] = distinct !{[[LOOP0]], [[META1:![0-9]+]], [[META2:![0-9]+]]}
|
|
; CHECK: [[META1]] = !{!"llvm.loop.isvectorized", i32 1}
|
|
; CHECK: [[META2]] = !{!"llvm.loop.unroll.runtime.disable"}
|
|
; CHECK: [[LOOP3]] = distinct !{[[LOOP3]], [[META2]], [[META1]]}
|
|
; CHECK: [[LOOP4]] = distinct !{[[LOOP4]], [[META1]], [[META2]]}
|
|
; CHECK: [[LOOP5]] = distinct !{[[LOOP5]], [[META2]], [[META1]]}
|
|
; CHECK: [[META6]] = !{[[META7:![0-9]+]]}
|
|
; CHECK: [[META7]] = distinct !{[[META7]], [[META8:![0-9]+]]}
|
|
; CHECK: [[META8]] = distinct !{[[META8]], !"LVerDomain"}
|
|
; CHECK: [[META9]] = !{[[META10:![0-9]+]]}
|
|
; CHECK: [[META10]] = distinct !{[[META10]], [[META8]]}
|
|
; CHECK: [[LOOP11]] = distinct !{[[LOOP11]], [[META1]], [[META2]]}
|
|
; CHECK: [[LOOP12]] = distinct !{[[LOOP12]], [[META1]]}
|
|
;.
|