Skip to content

Commit

Permalink
InstCombine: Fold is.fpclass(x, fcZero) to fcmp oeq 0
Browse files Browse the repository at this point in the history 
This requires the denormal mode to definitively be IEEE handling.
  • Loading branch information
@arsenm
arsenm committed Feb 15, 2023
1 parent 8f7e740 commit df78976
Show file tree
Hide file tree
Showing 2 changed files with 56 additions and 15 deletions.
41 changes: 41 additions & 0 deletions llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
View file
Expand Up @@ -830,11 +830,30 @@ InstCombinerImpl::foldIntrinsicWithOverflowCommon(IntrinsicInst *II) {
return nullptr;
}

/// \returns true if the test performed by llvm.is.fpclass(x, \p Mask) is
/// equivalent to fcmp oeq x, 0.0 with the floating-point environment assumed
/// for \p F for type \p Ty
static bool fpclassTestIsFCmp0(FPClassTest Mask, const Function &F, Type *Ty) {
if (Mask == fcZero)
return F.getDenormalMode(Ty->getScalarType()->getFltSemantics()).Input ==
DenormalMode::IEEE;

if (Mask == (fcZero | fcSubnormal)) {
DenormalMode::DenormalModeKind InputMode =
F.getDenormalMode(Ty->getScalarType()->getFltSemantics()).Input;
return InputMode == DenormalMode::PreserveSign ||
InputMode == DenormalMode::PositiveZero;
}

return false;
}

Instruction *InstCombinerImpl::foldIntrinsicIsFPClass(IntrinsicInst &II) {
Value *Src0 = II.getArgOperand(0);
Value *Src1 = II.getArgOperand(1);
const ConstantInt *CMask = cast<ConstantInt>(Src1);
uint32_t Mask = CMask->getZExtValue();
uint32_t InvertedMask = ~CMask->getZExtValue() & fcAllFlags;
const bool IsStrict = II.isStrictFP();

Value *FNegSrc;
Expand Down Expand Up @@ -895,6 +914,28 @@ Instruction *InstCombinerImpl::foldIntrinsicIsFPClass(IntrinsicInst &II) {
return replaceInstUsesWith(II, FCmp);
}

if (!IsStrict &&
fpclassTestIsFCmp0(static_cast<FPClassTest>(Mask),
*II.getParent()->getParent(), Src0->getType())) {
// Equivalent of == 0.
Value *FCmp =
Builder.CreateFCmpOEQ(Src0, ConstantFP::get(Src0->getType(), 0.0));

FCmp->takeName(&II);
return replaceInstUsesWith(II, FCmp);
}

if (!IsStrict &&
fpclassTestIsFCmp0(static_cast<FPClassTest>(InvertedMask),
*II.getParent()->getParent(), Src0->getType())) {
// Equivalent of !(x == 0).
Value *FCmp =
Builder.CreateFCmpUNE(Src0, ConstantFP::get(Src0->getType(), 0.0));

FCmp->takeName(&II);
return replaceInstUsesWith(II, FCmp);
}

// fp_class (nnan x), qnan|snan|other -> fp_class (nnan x), other
if ((Mask & fcNan) && isKnownNeverNaN(Src0, &getTargetLibraryInfo())) {
II.setArgOperand(1, ConstantInt::get(Src1->getType(), Mask & ~fcNan));
Expand Down
30 changes: 15 additions & 15 deletions llvm/test/Transforms/InstCombine/is_fpclass.ll
View file
Expand Up @@ -116,7 +116,7 @@ define i1 @test_class_isnan_f32_strict(float %x) {

define i1 @test_class_is_p0_n0_f32(float %x) {
; CHECK-LABEL: @test_class_is_p0_n0_f32(
; CHECK-NEXT: [[VAL:%.*]] = call i1 @llvm.is.fpclass.f32(float [[X:%.*]], i32 96)
; CHECK-NEXT: [[VAL:%.*]] = fcmp oeq float [[X:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[VAL]]
;
%val = call i1 @llvm.is.fpclass.f32(float %x, i32 96) ; fcZero
Expand All @@ -125,7 +125,7 @@ define i1 @test_class_is_p0_n0_f32(float %x) {

define <2 x i1> @test_class_is_p0_n0_v2f32(<2 x float> %x) {
; CHECK-LABEL: @test_class_is_p0_n0_v2f32(
; CHECK-NEXT: [[VAL:%.*]] = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> [[X:%.*]], i32 96)
; CHECK-NEXT: [[VAL:%.*]] = fcmp oeq <2 x float> [[X:%.*]], zeroinitializer
; CHECK-NEXT: ret <2 x i1> [[VAL]]
;
%val = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %x, i32 96) ; fcZero
Expand Down Expand Up @@ -170,7 +170,7 @@ define <2 x i1> @test_class_is_p0_n0_or_nan_v2f32(<2 x float> %x) {

define i1 @test_class_is_not_p0_n0_f32(float %x) {
; CHECK-LABEL: @test_class_is_not_p0_n0_f32(
; CHECK-NEXT: [[VAL:%.*]] = call i1 @llvm.is.fpclass.f32(float [[X:%.*]], i32 927)
; CHECK-NEXT: [[VAL:%.*]] = fcmp une float [[X:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[VAL]]
;
%val = call i1 @llvm.is.fpclass.f32(float %x, i32 927) ; ~fcZero & fcAllFlags
Expand All @@ -179,7 +179,7 @@ define i1 @test_class_is_not_p0_n0_f32(float %x) {

define <2 x i1> @test_class_is_not_p0_n0_v2f32(<2 x float> %x) {
; CHECK-LABEL: @test_class_is_not_p0_n0_v2f32(
; CHECK-NEXT: [[VAL:%.*]] = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> [[X:%.*]], i32 927)
; CHECK-NEXT: [[VAL:%.*]] = fcmp une <2 x float> [[X:%.*]], zeroinitializer
; CHECK-NEXT: ret <2 x i1> [[VAL]]
;
%val = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %x, i32 927) ; ~fcZero & fcAllFlags
Expand Down Expand Up @@ -215,7 +215,7 @@ define i1 @test_class_is_not_p0_n0_f32_dynamic(float %x) "denormal-fp-math"="iee

define i1 @test_class_is_not_p0_n0_psub_nsub_f32_daz(float %x) "denormal-fp-math"="ieee,preserve-sign" {
; CHECK-LABEL: @test_class_is_not_p0_n0_psub_nsub_f32_daz(
; CHECK-NEXT: [[VAL:%.*]] = call i1 @llvm.is.fpclass.f32(float [[X:%.*]], i32 783)
; CHECK-NEXT: [[VAL:%.*]] = fcmp une float [[X:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[VAL]]
;
%val = call i1 @llvm.is.fpclass.f32(float %x, i32 783) ; ~(fcZero|fcSubnormal) & fcAllFlags
Expand All @@ -224,7 +224,7 @@ define i1 @test_class_is_not_p0_n0_psub_nsub_f32_daz(float %x) "denormal-fp-math

define i1 @test_class_is_not_p0_n0_psub_nsub_f32_dapz(float %x) "denormal-fp-math"="ieee,positive-zero" {
; CHECK-LABEL: @test_class_is_not_p0_n0_psub_nsub_f32_dapz(
; CHECK-NEXT: [[VAL:%.*]] = call i1 @llvm.is.fpclass.f32(float [[X:%.*]], i32 783)
; CHECK-NEXT: [[VAL:%.*]] = fcmp une float [[X:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[VAL]]
;
%val = call i1 @llvm.is.fpclass.f32(float %x, i32 783) ; ~(fcZero|fcSubnormal) & fcAllFlags
Expand Down Expand Up @@ -278,7 +278,7 @@ define i1 @test_class_is_p0_n0_psub_nsub_f32(float %x) {

define i1 @test_class_is_p0_n0_psub_nsub_f32_daz(float %x) "denormal-fp-math"="ieee,preserve-sign" {
; CHECK-LABEL: @test_class_is_p0_n0_psub_nsub_f32_daz(
; CHECK-NEXT: [[VAL:%.*]] = call i1 @llvm.is.fpclass.f32(float [[X:%.*]], i32 240)
; CHECK-NEXT: [[VAL:%.*]] = fcmp oeq float [[X:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[VAL]]
;
%val = call i1 @llvm.is.fpclass.f32(float %x, i32 240) ; fcZero | fcSubnormal
Expand All @@ -287,7 +287,7 @@ define i1 @test_class_is_p0_n0_psub_nsub_f32_daz(float %x) "denormal-fp-math"="i

define i1 @test_class_is_p0_n0_psub_nsub_f32_dapz(float %x) "denormal-fp-math"="ieee,positive-zero" {
; CHECK-LABEL: @test_class_is_p0_n0_psub_nsub_f32_dapz(
; CHECK-NEXT: [[VAL:%.*]] = call i1 @llvm.is.fpclass.f32(float [[X:%.*]], i32 240)
; CHECK-NEXT: [[VAL:%.*]] = fcmp oeq float [[X:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[VAL]]
;
%val = call i1 @llvm.is.fpclass.f32(float %x, i32 240) ; fcZero | fcSubnormal
Expand All @@ -314,7 +314,7 @@ define <2 x i1> @test_class_is_p0_n0_psub_nsub_v2f32(<2 x float> %x) {

define <2 x i1> @test_class_is_p0_n0_psub_nsub_v2f32_daz(<2 x float> %x) "denormal-fp-math"="ieee,preserve-sign" {
; CHECK-LABEL: @test_class_is_p0_n0_psub_nsub_v2f32_daz(
; CHECK-NEXT: [[VAL:%.*]] = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> [[X:%.*]], i32 240)
; CHECK-NEXT: [[VAL:%.*]] = fcmp oeq <2 x float> [[X:%.*]], zeroinitializer
; CHECK-NEXT: ret <2 x i1> [[VAL]]
;
%val = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %x, i32 240) ; fcZero | fcSubnormal
Expand All @@ -323,7 +323,7 @@ define <2 x i1> @test_class_is_p0_n0_psub_nsub_v2f32_daz(<2 x float> %x) "denorm

define <2 x i1> @test_class_is_p0_n0_psub_nsub_v2f32_dapz(<2 x float> %x) "denormal-fp-math"="ieee,positive-zero" {
; CHECK-LABEL: @test_class_is_p0_n0_psub_nsub_v2f32_dapz(
; CHECK-NEXT: [[VAL:%.*]] = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> [[X:%.*]], i32 240)
; CHECK-NEXT: [[VAL:%.*]] = fcmp oeq <2 x float> [[X:%.*]], zeroinitializer
; CHECK-NEXT: ret <2 x i1> [[VAL]]
;
%val = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %x, i32 240) ; fcZero | fcSubnormal
Expand Down Expand Up @@ -1460,7 +1460,7 @@ define i1 @test_class_fneg_normal(float %arg) {
define i1 @test_class_fneg_zero(float %arg) {
;
; CHECK-LABEL: @test_class_fneg_zero(
; CHECK-NEXT: [[CLASS:%.*]] = call i1 @llvm.is.fpclass.f32(float [[ARG:%.*]], i32 96)
; CHECK-NEXT: [[CLASS:%.*]] = fcmp oeq float [[ARG:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[CLASS]]
;
%fneg = fneg float %arg
Expand Down Expand Up @@ -1715,7 +1715,7 @@ define i1 @test_class_fabs_negzero(float %arg) {
; -> poszero
define i1 @test_class_fabs_poszero(float %arg) {
; CHECK-LABEL: @test_class_fabs_poszero(
; CHECK-NEXT: [[CLASS:%.*]] = call i1 @llvm.is.fpclass.f32(float [[ARG:%.*]], i32 96)
; CHECK-NEXT: [[CLASS:%.*]] = fcmp oeq float [[ARG:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[CLASS]]
;
%fabs = call float @llvm.fabs.f32(float %arg)
Expand Down Expand Up @@ -1792,7 +1792,7 @@ define i1 @test_class_fabs_normal(float %arg) {
; -> poszero
define i1 @test_class_fabs_zero(float %arg) {
; CHECK-LABEL: @test_class_fabs_zero(
; CHECK-NEXT: [[CLASS:%.*]] = call i1 @llvm.is.fpclass.f32(float [[ARG:%.*]], i32 96)
; CHECK-NEXT: [[CLASS:%.*]] = fcmp oeq float [[ARG:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[CLASS]]
;
%fabs = call float @llvm.fabs.f32(float %arg)
Expand Down Expand Up @@ -2038,7 +2038,7 @@ define i1 @test_class_fneg_fabs_negsubnormal(float %arg) {

define i1 @test_class_fneg_fabs_negzero(float %arg) {
; CHECK-LABEL: @test_class_fneg_fabs_negzero(
; CHECK-NEXT: [[CLASS:%.*]] = call i1 @llvm.is.fpclass.f32(float [[ARG:%.*]], i32 96)
; CHECK-NEXT: [[CLASS:%.*]] = fcmp oeq float [[ARG:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[CLASS]]
;
%fabs = call float @llvm.fabs.f32(float %arg)
Expand Down Expand Up @@ -2122,7 +2122,7 @@ define i1 @test_class_fneg_fabs_normal(float %arg) {

define i1 @test_class_fneg_fabs_zero(float %arg) {
; CHECK-LABEL: @test_class_fneg_fabs_zero(
; CHECK-NEXT: [[CLASS:%.*]] = call i1 @llvm.is.fpclass.f32(float [[ARG:%.*]], i32 96)
; CHECK-NEXT: [[CLASS:%.*]] = fcmp oeq float [[ARG:%.*]], 0.000000e+00
; CHECK-NEXT: ret i1 [[CLASS]]
;
%fabs = call float @llvm.fabs.f32(float %arg)
Expand Down

0 comments on commit df78976

Please sign in to comment.