Skip to content

Commit

Permalink
AMDGPU: Use generic is.fpclass enum instead of locally defined copy
Browse files Browse the repository at this point in the history 
The generic intrinsic uses the same bitlayout as the amdgcn intrinsic,
so re-use the enum.
  • Loading branch information
@arsenm
arsenm committed Nov 11, 2022
1 parent 2082188 commit 8ea3cf4
Showing 1 changed file with 19 additions and 36 deletions.
55 changes: 19 additions & 36 deletions llvm/lib/Target/AMDGPU/AMDGPUInstCombineIntrinsic.cpp
View file
Expand Up @@ -17,6 +17,7 @@
#include "AMDGPUInstrInfo.h"
#include "AMDGPUTargetTransformInfo.h"
#include "GCNSubtarget.h"
#include "llvm/ADT/FloatingPointMode.h"
#include "llvm/IR/IntrinsicsAMDGPU.h"
#include "llvm/Transforms/InstCombine/InstCombiner.h"

Expand Down Expand Up @@ -417,23 +418,6 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
break;
}
case Intrinsic::amdgcn_class: {
enum {
S_NAN = 1 << 0, // Signaling NaN
Q_NAN = 1 << 1, // Quiet NaN
N_INFINITY = 1 << 2, // Negative infinity
N_NORMAL = 1 << 3, // Negative normal
N_SUBNORMAL = 1 << 4, // Negative subnormal
N_ZERO = 1 << 5, // Negative zero
P_ZERO = 1 << 6, // Positive zero
P_SUBNORMAL = 1 << 7, // Positive subnormal
P_NORMAL = 1 << 8, // Positive normal
P_INFINITY = 1 << 9 // Positive infinity
};

const uint32_t FullMask = S_NAN | Q_NAN | N_INFINITY | N_NORMAL |
N_SUBNORMAL | N_ZERO | P_ZERO | P_SUBNORMAL |
P_NORMAL | P_INFINITY;

Value *Src0 = II.getArgOperand(0);
Value *Src1 = II.getArgOperand(1);
const ConstantInt *CMask = dyn_cast<ConstantInt>(Src1);
Expand All @@ -452,22 +436,22 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
uint32_t Mask = CMask->getZExtValue();

// If all tests are made, it doesn't matter what the value is.
if ((Mask & FullMask) == FullMask) {
if ((Mask & fcAllFlags) == fcAllFlags) {
return IC.replaceInstUsesWith(II, ConstantInt::get(II.getType(), true));
}

if ((Mask & FullMask) == 0) {
if ((Mask & fcAllFlags) == 0) {
return IC.replaceInstUsesWith(II, ConstantInt::get(II.getType(), false));
}

if (Mask == (S_NAN | Q_NAN)) {
if (Mask == fcNan) {
// Equivalent of isnan. Replace with standard fcmp.
Value *FCmp = IC.Builder.CreateFCmpUNO(Src0, Src0);
FCmp->takeName(&II);
return IC.replaceInstUsesWith(II, FCmp);
}

if (Mask == (N_ZERO | P_ZERO)) {
if (Mask == fcZero) {
// Equivalent of == 0.
Value *FCmp =
IC.Builder.CreateFCmpOEQ(Src0, ConstantFP::get(Src0->getType(), 0.0));
Expand All @@ -477,10 +461,9 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
}

// fp_class (nnan x), qnan|snan|other -> fp_class (nnan x), other
if (((Mask & S_NAN) || (Mask & Q_NAN)) &&
isKnownNeverNaN(Src0, &IC.getTargetLibraryInfo())) {
if ((Mask & fcNan) && isKnownNeverNaN(Src0, &IC.getTargetLibraryInfo())) {
return IC.replaceOperand(
II, 1, ConstantInt::get(Src1->getType(), Mask & ~(S_NAN | Q_NAN)));
II, 1, ConstantInt::get(Src1->getType(), Mask & ~fcNan));
}

const ConstantFP *CVal = dyn_cast<ConstantFP>(Src0);
Expand All @@ -490,10 +473,10 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
}

// Clamp mask to used bits
if ((Mask & FullMask) != Mask) {
if ((Mask & fcAllFlags) != Mask) {
CallInst *NewCall = IC.Builder.CreateCall(
II.getCalledFunction(),
{Src0, ConstantInt::get(Src1->getType(), Mask & FullMask)});
{Src0, ConstantInt::get(Src1->getType(), Mask & fcAllFlags)});

NewCall->takeName(&II);
return IC.replaceInstUsesWith(II, NewCall);
Expand All @@ -505,16 +488,16 @@ GCNTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
const APFloat &Val = CVal->getValueAPF();

bool Result =
((Mask & S_NAN) && Val.isNaN() && Val.isSignaling()) ||
((Mask & Q_NAN) && Val.isNaN() && !Val.isSignaling()) ||
((Mask & N_INFINITY) && Val.isInfinity() && Val.isNegative()) ||
((Mask & N_NORMAL) && Val.isNormal() && Val.isNegative()) ||
((Mask & N_SUBNORMAL) && Val.isDenormal() && Val.isNegative()) ||
((Mask & N_ZERO) && Val.isZero() && Val.isNegative()) ||
((Mask & P_ZERO) && Val.isZero() && !Val.isNegative()) ||
((Mask & P_SUBNORMAL) && Val.isDenormal() && !Val.isNegative()) ||
((Mask & P_NORMAL) && Val.isNormal() && !Val.isNegative()) ||
((Mask & P_INFINITY) && Val.isInfinity() && !Val.isNegative());
((Mask & fcSNan) && Val.isNaN() && Val.isSignaling()) ||
((Mask & fcQNan) && Val.isNaN() && !Val.isSignaling()) ||
((Mask & fcNegInf) && Val.isInfinity() && Val.isNegative()) ||
((Mask & fcNegNormal) && Val.isNormal() && Val.isNegative()) ||
((Mask & fcNegSubnormal) && Val.isDenormal() && Val.isNegative()) ||
((Mask & fcNegZero) && Val.isZero() && Val.isNegative()) ||
((Mask & fcPosZero) && Val.isZero() && !Val.isNegative()) ||
((Mask & fcPosSubnormal) && Val.isDenormal() && !Val.isNegative()) ||
((Mask & fcPosNormal) && Val.isNormal() && !Val.isNegative()) ||
((Mask & fcPosInf) && Val.isInfinity() && !Val.isNegative());

return IC.replaceInstUsesWith(II, ConstantInt::get(II.getType(), Result));
}
Expand Down

0 comments on commit 8ea3cf4

Please sign in to comment.