Skip to content

[X86] Support lowering of FMINIMUMNUM/FMAXIMUMNUM #121464

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Jan 6, 2025
Merged

[X86] Support lowering of FMINIMUMNUM/FMAXIMUMNUM #121464

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
398068e
[X86] Support lowering of FMINIMUMNUM/FMAXIMUMNUM
phoebewang Jan 2, 2025
204186c
Remove unnecessary check
phoebewang Jan 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -402,6 +402,8 @@ SDValue VectorLegalizer::LegalizeOp(SDValue Op) {
case ISD::FMAXNUM_IEEE:
case ISD::FMINIMUM:
case ISD::FMAXIMUM:
case ISD::FMINIMUMNUM:
case ISD::FMAXIMUMNUM:
case ISD::FCOPYSIGN:
case ISD::FSQRT:
case ISD::FSIN:
Expand Down Expand Up @@ -1081,6 +1083,10 @@ void VectorLegalizer::Expand(SDNode *Node, SmallVectorImpl<SDValue> &Results) {
case ISD::FMAXIMUM:
Results.push_back(TLI.expandFMINIMUM_FMAXIMUM(Node, DAG));
return;
case ISD::FMINIMUMNUM:
case ISD::FMAXIMUMNUM:
Results.push_back(TLI.expandFMINIMUMNUM_FMAXIMUMNUM(Node, DAG));
return;
case ISD::SMIN:
case ISD::SMAX:
case ISD::UMIN:
Expand Down
2 changes: 2 additions & 0 deletions llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -149,6 +149,8 @@ void DAGTypeLegalizer::ScalarizeVectorResult(SDNode *N, unsigned ResNo) {
case ISD::FMAXNUM_IEEE:
case ISD::FMINIMUM:
case ISD::FMAXIMUM:
case ISD::FMINIMUMNUM:
case ISD::FMAXIMUMNUM:
case ISD::FLDEXP:
case ISD::ABDS:
case ISD::ABDU:
Expand Down
40 changes: 35 additions & 5 deletions llvm/lib/Target/X86/X86ISelLowering.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -623,6 +623,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
setOperationAction(ISD::FMAXNUM, VT, Action);
setOperationAction(ISD::FMINIMUM, VT, Action);
setOperationAction(ISD::FMAXIMUM, VT, Action);
setOperationAction(ISD::FMINIMUMNUM, VT, Action);
setOperationAction(ISD::FMAXIMUMNUM, VT, Action);
setOperationAction(ISD::FSIN, VT, Action);
setOperationAction(ISD::FCOS, VT, Action);
setOperationAction(ISD::FSINCOS, VT, Action);
Expand Down Expand Up @@ -1066,6 +1068,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,

setOperationAction(ISD::FMAXIMUM, MVT::f32, Custom);
setOperationAction(ISD::FMINIMUM, MVT::f32, Custom);
setOperationAction(ISD::FMAXIMUMNUM, MVT::f32, Custom);
setOperationAction(ISD::FMINIMUMNUM, MVT::f32, Custom);

setOperationAction(ISD::FNEG, MVT::v4f32, Custom);
setOperationAction(ISD::FABS, MVT::v4f32, Custom);
Expand Down Expand Up @@ -1108,6 +1112,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
for (auto VT : { MVT::f64, MVT::v4f32, MVT::v2f64 }) {
setOperationAction(ISD::FMAXIMUM, VT, Custom);
setOperationAction(ISD::FMINIMUM, VT, Custom);
setOperationAction(ISD::FMAXIMUMNUM, VT, Custom);
setOperationAction(ISD::FMINIMUMNUM, VT, Custom);
}

for (auto VT : { MVT::v2i8, MVT::v4i8, MVT::v8i8,
Expand Down Expand Up @@ -1473,6 +1479,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,

setOperationAction(ISD::FMAXIMUM, VT, Custom);
setOperationAction(ISD::FMINIMUM, VT, Custom);
setOperationAction(ISD::FMAXIMUMNUM, VT, Custom);
setOperationAction(ISD::FMINIMUMNUM, VT, Custom);
setOperationAction(ISD::FCANONICALIZE, VT, Custom);
}

Expand Down Expand Up @@ -1818,6 +1826,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
for (MVT VT : { MVT::v16f32, MVT::v8f64 }) {
setOperationAction(ISD::FMAXIMUM, VT, Custom);
setOperationAction(ISD::FMINIMUM, VT, Custom);
setOperationAction(ISD::FMAXIMUMNUM, VT, Custom);
setOperationAction(ISD::FMINIMUMNUM, VT, Custom);
setOperationAction(ISD::FNEG, VT, Custom);
setOperationAction(ISD::FABS, VT, Custom);
setOperationAction(ISD::FMA, VT, Legal);
Expand Down Expand Up @@ -2289,6 +2299,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
setOperationAction(ISD::STRICT_FP_ROUND, MVT::f16, Custom);
setOperationAction(ISD::FMAXIMUM, MVT::f16, Custom);
setOperationAction(ISD::FMINIMUM, MVT::f16, Custom);
setOperationAction(ISD::FMAXIMUMNUM, MVT::f16, Custom);
setOperationAction(ISD::FMINIMUMNUM, MVT::f16, Custom);
setOperationAction(ISD::FP_EXTEND, MVT::f32, Legal);
setOperationAction(ISD::STRICT_FP_EXTEND, MVT::f32, Legal);

Expand Down Expand Up @@ -2336,6 +2348,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,

setOperationAction(ISD::FMINIMUM, MVT::v32f16, Custom);
setOperationAction(ISD::FMAXIMUM, MVT::v32f16, Custom);
setOperationAction(ISD::FMINIMUMNUM, MVT::v32f16, Custom);
setOperationAction(ISD::FMAXIMUMNUM, MVT::v32f16, Custom);
}

if (Subtarget.hasVLX()) {
Expand Down Expand Up @@ -2383,9 +2397,13 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,

setOperationAction(ISD::FMINIMUM, MVT::v8f16, Custom);
setOperationAction(ISD::FMAXIMUM, MVT::v8f16, Custom);
setOperationAction(ISD::FMINIMUMNUM, MVT::v8f16, Custom);
setOperationAction(ISD::FMAXIMUMNUM, MVT::v8f16, Custom);

setOperationAction(ISD::FMINIMUM, MVT::v16f16, Custom);
setOperationAction(ISD::FMAXIMUM, MVT::v16f16, Custom);
setOperationAction(ISD::FMINIMUMNUM, MVT::v16f16, Custom);
setOperationAction(ISD::FMAXIMUMNUM, MVT::v16f16, Custom);
}
}

Expand Down Expand Up @@ -2444,6 +2462,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
setOperationAction(ISD::SETCC, VT, Custom);
setOperationAction(ISD::FMINIMUM, VT, Custom);
setOperationAction(ISD::FMAXIMUM, VT, Custom);
setOperationAction(ISD::FMINIMUMNUM, VT, Custom);
setOperationAction(ISD::FMAXIMUMNUM, VT, Custom);
}
if (Subtarget.hasAVX10_2_512()) {
setOperationAction(ISD::FADD, MVT::v32bf16, Legal);
Expand All @@ -2455,6 +2475,8 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
setOperationAction(ISD::SETCC, MVT::v32bf16, Custom);
setOperationAction(ISD::FMINIMUM, MVT::v32bf16, Custom);
setOperationAction(ISD::FMAXIMUM, MVT::v32bf16, Custom);
setOperationAction(ISD::FMINIMUMNUM, MVT::v32bf16, Custom);
setOperationAction(ISD::FMAXIMUMNUM, MVT::v32bf16, Custom);
}
for (auto VT : {MVT::f16, MVT::f32, MVT::f64}) {
setCondCodeAction(ISD::SETOEQ, VT, Custom);
Expand Down Expand Up @@ -28839,13 +28861,15 @@ static SDValue LowerMINMAX(SDValue Op, const X86Subtarget &Subtarget,

static SDValue LowerFMINIMUM_FMAXIMUM(SDValue Op, const X86Subtarget &Subtarget,
SelectionDAG &DAG) {
assert((Op.getOpcode() == ISD::FMAXIMUM || Op.getOpcode() == ISD::FMINIMUM) &&
"Expected FMAXIMUM or FMINIMUM opcode");
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
EVT VT = Op.getValueType();
SDValue X = Op.getOperand(0);
SDValue Y = Op.getOperand(1);
SDLoc DL(Op);
bool IsMaxOp =
Op.getOpcode() == ISD::FMAXIMUM || Op.getOpcode() == ISD::FMAXIMUMNUM;
bool IsNum =
Op.getOpcode() == ISD::FMINIMUMNUM || Op.getOpcode() == ISD::FMAXIMUMNUM;
if (Subtarget.hasAVX10_2() && TLI.isTypeLegal(VT)) {
unsigned Opc = 0;
if (VT.isVector())
Expand All @@ -28855,7 +28879,7 @@ static SDValue LowerFMINIMUM_FMAXIMUM(SDValue Op, const X86Subtarget &Subtarget,

if (Opc) {
SDValue Imm =
DAG.getTargetConstant(Op.getOpcode() == ISD::FMAXIMUM, DL, MVT::i32);
DAG.getTargetConstant(IsMaxOp + (IsNum ? 16 : 0), DL, MVT::i32);
return DAG.getNode(Opc, DL, VT, X, Y, Imm, Op->getFlags());
}
}
Expand All @@ -28865,7 +28889,7 @@ static SDValue LowerFMINIMUM_FMAXIMUM(SDValue Op, const X86Subtarget &Subtarget,
APInt OppositeZero = PreferredZero;
EVT IVT = VT.changeTypeToInteger();
X86ISD::NodeType MinMaxOp;
if (Op.getOpcode() == ISD::FMAXIMUM) {
if (IsMaxOp) {
MinMaxOp = X86ISD::FMAX;
OppositeZero.setSignBit();
} else {
Expand Down Expand Up @@ -28995,7 +29019,9 @@ static SDValue LowerFMINIMUM_FMAXIMUM(SDValue Op, const X86Subtarget &Subtarget,
if (IgnoreNaN || DAG.isKnownNeverNaN(NewX))
return MinMax;

SDValue IsNaN = DAG.getSetCC(DL, SetCCType, NewX, NewX, ISD::SETUO);
SDValue IsNaN =
DAG.getSetCC(DL, SetCCType, NewX, NewX, IsNum ? ISD::SETO : ISD::SETUO);

return DAG.getSelect(DL, VT, IsNaN, NewX, MinMax);
}

Expand Down Expand Up @@ -33253,6 +33279,8 @@ SDValue X86TargetLowering::LowerOperation(SDValue Op, SelectionDAG &DAG) const {
case ISD::UMIN: return LowerMINMAX(Op, Subtarget, DAG);
case ISD::FMINIMUM:
case ISD::FMAXIMUM:
case ISD::FMINIMUMNUM:
case ISD::FMAXIMUMNUM:
return LowerFMINIMUM_FMAXIMUM(Op, Subtarget, DAG);
case ISD::ABS: return LowerABS(Op, Subtarget, DAG);
case ISD::ABDS:
Expand Down Expand Up @@ -45994,6 +46022,8 @@ static SDValue scalarizeExtEltFP(SDNode *ExtElt, SelectionDAG &DAG,
case ISD::FMAXNUM_IEEE:
case ISD::FMAXIMUM:
case ISD::FMINIMUM:
case ISD::FMAXIMUMNUM:
case ISD::FMINIMUMNUM:
case X86ISD::FMAX:
case X86ISD::FMIN:
case ISD::FABS: // Begin 1 operand
Expand Down
120 changes: 60 additions & 60 deletions llvm/test/CodeGen/AMDGPU/maximumnum.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1838,11 +1838,11 @@ define <3 x half> @v_maximumnum_v3f16(<3 x half> %x, <3 x half> %y) {
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v4, v5, v4 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_max_f16_e32 v2, v3, v3
; GFX8-NEXT: v_max_f16_e32 v3, v3, v3
; GFX8-NEXT: v_max_f16_e32 v1, v1, v1
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_max_f16_e32 v1, v1, v3
; GFX8-NEXT: v_or_b32_e32 v0, v0, v4
; GFX8-NEXT: v_max_f16_e32 v1, v1, v2
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_maximumnum_v3f16:
Expand Down Expand Up @@ -1904,8 +1904,8 @@ define <3 x half> @v_maximumnum_v3f16_nnan(<3 x half> %x, <3 x half> %y) {
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_max_f16_sdwa v4, v0, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_or_b32_e32 v0, v0, v4
; GFX8-NEXT: v_max_f16_e32 v1, v1, v3
; GFX8-NEXT: v_or_b32_e32 v0, v0, v4
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_maximumnum_v3f16_nnan:
Expand Down Expand Up @@ -1947,20 +1947,20 @@ define <4 x half> @v_maximumnum_v4f16(<4 x half> %x, <4 x half> %y) {
; GFX8-LABEL: v_maximumnum_v4f16:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_max_f16_sdwa v4, v2, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v5, v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v4, v3, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v5, v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v4, v5, v4 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_or_b32_e32 v0, v0, v4
; GFX8-NEXT: v_max_f16_sdwa v2, v3, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v4, v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v5, v2, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v6, v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v3, v3, v3
; GFX8-NEXT: v_max_f16_e32 v1, v1, v1
; GFX8-NEXT: v_max_f16_sdwa v2, v4, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v5, v6, v5 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v1, v1, v3
; GFX8-NEXT: v_or_b32_e32 v1, v1, v2
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_or_b32_e32 v0, v0, v5
; GFX8-NEXT: v_or_b32_e32 v1, v1, v4
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_maximumnum_v4f16:
Expand Down Expand Up @@ -2020,12 +2020,12 @@ define <4 x half> @v_maximumnum_v4f16_nnan(<4 x half> %x, <4 x half> %y) {
; GFX8-LABEL: v_maximumnum_v4f16_nnan:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_max_f16_sdwa v4, v0, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_max_f16_sdwa v2, v1, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v4, v1, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v5, v0, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v1, v1, v3
; GFX8-NEXT: v_or_b32_e32 v0, v0, v4
; GFX8-NEXT: v_or_b32_e32 v1, v1, v2
; GFX8-NEXT: v_max_f16_e32 v0, v0, v2
; GFX8-NEXT: v_or_b32_e32 v0, v0, v5
; GFX8-NEXT: v_or_b32_e32 v1, v1, v4
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_maximumnum_v4f16_nnan:
Expand Down Expand Up @@ -2067,27 +2067,27 @@ define <6 x half> @v_maximumnum_v6f16(<6 x half> %x, <6 x half> %y) {
; GFX8-LABEL: v_maximumnum_v6f16:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_max_f16_sdwa v6, v3, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v7, v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v3, v3, v3
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v6, v5, v5 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v7, v2, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v6, v7, v6 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v0, v0, v3
; GFX8-NEXT: v_or_b32_e32 v0, v0, v6
; GFX8-NEXT: v_max_f16_sdwa v3, v4, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v6, v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v7, v4, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v8, v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v7, v8, v7 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_sdwa v8, v3, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v9, v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v5, v5, v5
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v4, v4, v4
; GFX8-NEXT: v_max_f16_e32 v1, v1, v1
; GFX8-NEXT: v_max_f16_sdwa v3, v6, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v3, v3, v3
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v8, v9, v8 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v2, v2, v5
; GFX8-NEXT: v_max_f16_e32 v1, v1, v4
; GFX8-NEXT: v_or_b32_e32 v1, v1, v3
; GFX8-NEXT: v_max_f16_sdwa v3, v5, v5 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v4, v2, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v3, v4, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v4, v5, v5
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v2, v2, v4
; GFX8-NEXT: v_or_b32_e32 v2, v2, v3
; GFX8-NEXT: v_max_f16_e32 v0, v0, v3
; GFX8-NEXT: v_or_b32_e32 v0, v0, v8
; GFX8-NEXT: v_or_b32_e32 v1, v1, v7
; GFX8-NEXT: v_or_b32_e32 v2, v2, v6
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_maximumnum_v6f16:
Expand Down Expand Up @@ -2159,34 +2159,34 @@ define <8 x half> @v_maximumnum_v8f16(<8 x half> %x, <8 x half> %y) {
; GFX8-LABEL: v_maximumnum_v8f16:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_max_f16_sdwa v8, v4, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v9, v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v4, v4, v4
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v8, v7, v7 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v9, v3, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v8, v9, v8 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v0, v0, v4
; GFX8-NEXT: v_or_b32_e32 v0, v0, v8
; GFX8-NEXT: v_max_f16_sdwa v4, v5, v5 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v8, v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v9, v6, v6 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v10, v2, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v9, v10, v9 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_sdwa v10, v5, v5 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v11, v1, v1 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v10, v11, v10 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_sdwa v11, v4, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v12, v0, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_e32 v7, v7, v7
; GFX8-NEXT: v_max_f16_e32 v3, v3, v3
; GFX8-NEXT: v_max_f16_e32 v6, v6, v6
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v5, v5, v5
; GFX8-NEXT: v_max_f16_e32 v1, v1, v1
; GFX8-NEXT: v_max_f16_sdwa v4, v8, v4 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v4, v4, v4
; GFX8-NEXT: v_max_f16_e32 v0, v0, v0
; GFX8-NEXT: v_max_f16_sdwa v11, v12, v11 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v3, v3, v7
; GFX8-NEXT: v_max_f16_e32 v2, v2, v6
; GFX8-NEXT: v_max_f16_e32 v1, v1, v5
; GFX8-NEXT: v_or_b32_e32 v1, v1, v4
; GFX8-NEXT: v_max_f16_sdwa v4, v6, v6 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v5, v2, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v4, v5, v4 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v5, v6, v6
; GFX8-NEXT: v_max_f16_e32 v2, v2, v2
; GFX8-NEXT: v_max_f16_e32 v2, v2, v5
; GFX8-NEXT: v_or_b32_e32 v2, v2, v4
; GFX8-NEXT: v_max_f16_sdwa v4, v7, v7 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v5, v3, v3 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_max_f16_sdwa v4, v5, v4 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
; GFX8-NEXT: v_max_f16_e32 v5, v7, v7
; GFX8-NEXT: v_max_f16_e32 v3, v3, v3
; GFX8-NEXT: v_max_f16_e32 v3, v3, v5
; GFX8-NEXT: v_or_b32_e32 v3, v3, v4
; GFX8-NEXT: v_max_f16_e32 v0, v0, v4
; GFX8-NEXT: v_or_b32_e32 v0, v0, v11
; GFX8-NEXT: v_or_b32_e32 v1, v1, v10
; GFX8-NEXT: v_or_b32_e32 v2, v2, v9
; GFX8-NEXT: v_or_b32_e32 v3, v3, v8
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_maximumnum_v8f16:
Expand Down
Loading
Loading