9529597cf4
This was reverted because of a miscompilation. At closer inspection, the problem was actually visible in a changed llvm regression test too. This one-line follow up fix/recommit will splat the IV, which is what we are trying to avoid if unnecessary in general, if tail-folding is requested even if all users are scalar instructions after vectorisation. Because with tail-folding, the splat IV will be used by the predicate of the masked loads/stores instructions. The previous version omitted this, which caused the miscompilation. The original commit message was: If tail-folding of the scalar remainder loop is applied, the primary induction variable is splat to a vector and used by the masked load/store vector instructions, thus the IV does not remain scalar. Because we now mark that the IV does not remain scalar for these cases, we don't emit the vector IV if it is not used. Thus, the vectoriser produces less dead code. Thanks to Ayal Zaks for the direction how to fix this.
175 lines
7.8 KiB
LLVM
175 lines
7.8 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
|
|
; RUN: opt -S -loop-vectorize -mtriple=x86_64-apple-darwin %s | FileCheck %s --check-prefixes=CHECK,SSE
|
|
; RUN: opt -S -loop-vectorize -mtriple=x86_64-apple-darwin -mattr=+avx %s | FileCheck %s --check-prefixes=CHECK,AVX
|
|
|
|
; Two mostly identical functions. The only difference is the presence of
|
|
; fast-math flags on the second. The loop is a pretty simple reduction:
|
|
|
|
; for (int i = 0; i < 32; ++i)
|
|
; if (arr[i] != 42)
|
|
; tot += arr[i];
|
|
|
|
define double @sumIfScalar(double* nocapture readonly %arr) {
|
|
; CHECK-LABEL: @sumIfScalar(
|
|
; CHECK-NEXT: entry:
|
|
; CHECK-NEXT: br label [[LOOP:%.*]]
|
|
; CHECK: loop:
|
|
; CHECK-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[NEXT_ITER:%.*]] ]
|
|
; CHECK-NEXT: [[TOT:%.*]] = phi double [ 0.000000e+00, [[ENTRY]] ], [ [[TOT_NEXT:%.*]], [[NEXT_ITER]] ]
|
|
; CHECK-NEXT: [[ADDR:%.*]] = getelementptr double, double* [[ARR:%.*]], i32 [[I]]
|
|
; CHECK-NEXT: [[NEXTVAL:%.*]] = load double, double* [[ADDR]]
|
|
; CHECK-NEXT: [[TST:%.*]] = fcmp une double [[NEXTVAL]], 4.200000e+01
|
|
; CHECK-NEXT: br i1 [[TST]], label [[DO_ADD:%.*]], label [[NO_ADD:%.*]]
|
|
; CHECK: do.add:
|
|
; CHECK-NEXT: [[TOT_NEW:%.*]] = fadd double [[TOT]], [[NEXTVAL]]
|
|
; CHECK-NEXT: br label [[NEXT_ITER]]
|
|
; CHECK: no.add:
|
|
; CHECK-NEXT: br label [[NEXT_ITER]]
|
|
; CHECK: next.iter:
|
|
; CHECK-NEXT: [[TOT_NEXT]] = phi double [ [[TOT]], [[NO_ADD]] ], [ [[TOT_NEW]], [[DO_ADD]] ]
|
|
; CHECK-NEXT: [[I_NEXT]] = add i32 [[I]], 1
|
|
; CHECK-NEXT: [[AGAIN:%.*]] = icmp ult i32 [[I_NEXT]], 32
|
|
; CHECK-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE:%.*]]
|
|
; CHECK: done:
|
|
; CHECK-NEXT: [[TOT_NEXT_LCSSA:%.*]] = phi double [ [[TOT_NEXT]], [[NEXT_ITER]] ]
|
|
; CHECK-NEXT: ret double [[TOT_NEXT_LCSSA]]
|
|
;
|
|
entry:
|
|
br label %loop
|
|
|
|
loop:
|
|
%i = phi i32 [0, %entry], [%i.next, %next.iter]
|
|
%tot = phi double [0.0, %entry], [%tot.next, %next.iter]
|
|
|
|
%addr = getelementptr double, double* %arr, i32 %i
|
|
%nextval = load double, double* %addr
|
|
|
|
%tst = fcmp une double %nextval, 42.0
|
|
br i1 %tst, label %do.add, label %no.add
|
|
|
|
do.add:
|
|
%tot.new = fadd double %tot, %nextval
|
|
br label %next.iter
|
|
|
|
no.add:
|
|
br label %next.iter
|
|
|
|
next.iter:
|
|
%tot.next = phi double [%tot, %no.add], [%tot.new, %do.add]
|
|
%i.next = add i32 %i, 1
|
|
%again = icmp ult i32 %i.next, 32
|
|
br i1 %again, label %loop, label %done
|
|
|
|
done:
|
|
ret double %tot.next
|
|
}
|
|
|
|
define double @sumIfVector(double* nocapture readonly %arr) {
|
|
; SSE-LABEL: @sumIfVector(
|
|
; SSE-NEXT: entry:
|
|
; SSE-NEXT: br label [[LOOP:%.*]]
|
|
; SSE: loop:
|
|
; SSE-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[NEXT_ITER:%.*]] ]
|
|
; SSE-NEXT: [[TOT:%.*]] = phi double [ 0.000000e+00, [[ENTRY]] ], [ [[TOT_NEXT:%.*]], [[NEXT_ITER]] ]
|
|
; SSE-NEXT: [[ADDR:%.*]] = getelementptr double, double* [[ARR:%.*]], i32 [[I]]
|
|
; SSE-NEXT: [[NEXTVAL:%.*]] = load double, double* [[ADDR]]
|
|
; SSE-NEXT: [[TST:%.*]] = fcmp fast une double [[NEXTVAL]], 4.200000e+01
|
|
; SSE-NEXT: br i1 [[TST]], label [[DO_ADD:%.*]], label [[NO_ADD:%.*]]
|
|
; SSE: do.add:
|
|
; SSE-NEXT: [[TOT_NEW:%.*]] = fadd fast double [[TOT]], [[NEXTVAL]]
|
|
; SSE-NEXT: br label [[NEXT_ITER]]
|
|
; SSE: no.add:
|
|
; SSE-NEXT: br label [[NEXT_ITER]]
|
|
; SSE: next.iter:
|
|
; SSE-NEXT: [[TOT_NEXT]] = phi double [ [[TOT]], [[NO_ADD]] ], [ [[TOT_NEW]], [[DO_ADD]] ]
|
|
; SSE-NEXT: [[I_NEXT]] = add i32 [[I]], 1
|
|
; SSE-NEXT: [[AGAIN:%.*]] = icmp ult i32 [[I_NEXT]], 32
|
|
; SSE-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE:%.*]]
|
|
; SSE: done:
|
|
; SSE-NEXT: [[TOT_NEXT_LCSSA:%.*]] = phi double [ [[TOT_NEXT]], [[NEXT_ITER]] ]
|
|
; SSE-NEXT: ret double [[TOT_NEXT_LCSSA]]
|
|
;
|
|
; AVX-LABEL: @sumIfVector(
|
|
; AVX-NEXT: entry:
|
|
; AVX-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
|
|
; AVX: vector.ph:
|
|
; AVX-NEXT: br label [[VECTOR_BODY:%.*]]
|
|
; AVX: vector.body:
|
|
; AVX-NEXT: [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
|
|
; AVX-NEXT: [[VEC_PHI:%.*]] = phi <4 x double> [ zeroinitializer, [[VECTOR_PH]] ], [ [[PREDPHI:%.*]], [[VECTOR_BODY]] ]
|
|
; AVX-NEXT: [[TMP0:%.*]] = add i32 [[INDEX]], 0
|
|
; AVX-NEXT: [[TMP1:%.*]] = getelementptr double, double* [[ARR:%.*]], i32 [[TMP0]]
|
|
; AVX-NEXT: [[TMP2:%.*]] = getelementptr double, double* [[TMP1]], i32 0
|
|
; AVX-NEXT: [[TMP3:%.*]] = bitcast double* [[TMP2]] to <4 x double>*
|
|
; AVX-NEXT: [[WIDE_LOAD:%.*]] = load <4 x double>, <4 x double>* [[TMP3]], align 8
|
|
; AVX-NEXT: [[TMP4:%.*]] = fcmp fast une <4 x double> [[WIDE_LOAD]], <double 4.200000e+01, double 4.200000e+01, double 4.200000e+01, double 4.200000e+01>
|
|
; AVX-NEXT: [[TMP5:%.*]] = fadd fast <4 x double> [[VEC_PHI]], [[WIDE_LOAD]]
|
|
; AVX-NEXT: [[TMP6:%.*]] = xor <4 x i1> [[TMP4]], <i1 true, i1 true, i1 true, i1 true>
|
|
; AVX-NEXT: [[PREDPHI]] = select <4 x i1> [[TMP4]], <4 x double> [[TMP5]], <4 x double> [[VEC_PHI]]
|
|
; AVX-NEXT: [[INDEX_NEXT]] = add i32 [[INDEX]], 4
|
|
; AVX-NEXT: [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 32
|
|
; AVX-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop !0
|
|
; AVX: middle.block:
|
|
; AVX-NEXT: [[RDX_SHUF:%.*]] = shufflevector <4 x double> [[PREDPHI]], <4 x double> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
|
|
; AVX-NEXT: [[BIN_RDX:%.*]] = fadd fast <4 x double> [[PREDPHI]], [[RDX_SHUF]]
|
|
; AVX-NEXT: [[RDX_SHUF1:%.*]] = shufflevector <4 x double> [[BIN_RDX]], <4 x double> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
|
|
; AVX-NEXT: [[BIN_RDX2:%.*]] = fadd fast <4 x double> [[BIN_RDX]], [[RDX_SHUF1]]
|
|
; AVX-NEXT: [[TMP8:%.*]] = extractelement <4 x double> [[BIN_RDX2]], i32 0
|
|
; AVX-NEXT: [[CMP_N:%.*]] = icmp eq i32 32, 32
|
|
; AVX-NEXT: br i1 [[CMP_N]], label [[DONE:%.*]], label [[SCALAR_PH]]
|
|
; AVX: scalar.ph:
|
|
; AVX-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ 32, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
|
|
; AVX-NEXT: [[BC_MERGE_RDX:%.*]] = phi double [ 0.000000e+00, [[ENTRY]] ], [ [[TMP8]], [[MIDDLE_BLOCK]] ]
|
|
; AVX-NEXT: br label [[LOOP:%.*]]
|
|
; AVX: loop:
|
|
; AVX-NEXT: [[I:%.*]] = phi i32 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[I_NEXT:%.*]], [[NEXT_ITER:%.*]] ]
|
|
; AVX-NEXT: [[TOT:%.*]] = phi double [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TOT_NEXT:%.*]], [[NEXT_ITER]] ]
|
|
; AVX-NEXT: [[ADDR:%.*]] = getelementptr double, double* [[ARR]], i32 [[I]]
|
|
; AVX-NEXT: [[NEXTVAL:%.*]] = load double, double* [[ADDR]]
|
|
; AVX-NEXT: [[TST:%.*]] = fcmp fast une double [[NEXTVAL]], 4.200000e+01
|
|
; AVX-NEXT: br i1 [[TST]], label [[DO_ADD:%.*]], label [[NO_ADD:%.*]]
|
|
; AVX: do.add:
|
|
; AVX-NEXT: [[TOT_NEW:%.*]] = fadd fast double [[TOT]], [[NEXTVAL]]
|
|
; AVX-NEXT: br label [[NEXT_ITER]]
|
|
; AVX: no.add:
|
|
; AVX-NEXT: br label [[NEXT_ITER]]
|
|
; AVX: next.iter:
|
|
; AVX-NEXT: [[TOT_NEXT]] = phi double [ [[TOT]], [[NO_ADD]] ], [ [[TOT_NEW]], [[DO_ADD]] ]
|
|
; AVX-NEXT: [[I_NEXT]] = add i32 [[I]], 1
|
|
; AVX-NEXT: [[AGAIN:%.*]] = icmp ult i32 [[I_NEXT]], 32
|
|
; AVX-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE]], !llvm.loop !2
|
|
; AVX: done:
|
|
; AVX-NEXT: [[TOT_NEXT_LCSSA:%.*]] = phi double [ [[TOT_NEXT]], [[NEXT_ITER]] ], [ [[TMP8]], [[MIDDLE_BLOCK]] ]
|
|
; AVX-NEXT: ret double [[TOT_NEXT_LCSSA]]
|
|
;
|
|
entry:
|
|
br label %loop
|
|
|
|
loop:
|
|
%i = phi i32 [0, %entry], [%i.next, %next.iter]
|
|
%tot = phi double [0.0, %entry], [%tot.next, %next.iter]
|
|
|
|
%addr = getelementptr double, double* %arr, i32 %i
|
|
%nextval = load double, double* %addr
|
|
|
|
%tst = fcmp fast une double %nextval, 42.0
|
|
br i1 %tst, label %do.add, label %no.add
|
|
|
|
do.add:
|
|
%tot.new = fadd fast double %tot, %nextval
|
|
br label %next.iter
|
|
|
|
no.add:
|
|
br label %next.iter
|
|
|
|
next.iter:
|
|
%tot.next = phi double [%tot, %no.add], [%tot.new, %do.add]
|
|
%i.next = add i32 %i, 1
|
|
%again = icmp ult i32 %i.next, 32
|
|
br i1 %again, label %loop, label %done
|
|
|
|
done:
|
|
ret double %tot.next
|
|
}
|
|
|