[AArch64] Add ARMv8.2-A FP16 scalar instructions
ARMv8.2-A adds 16-bit floating point versions of all existing VFP floating-point instructions. This is an optional extension, so all of these instructions require the FeatureFullFP16 subtarget feature. Most of these instructions are the same as the 32- and 64-bit versions, but with the type field (bits 23-22) set to 0b11. Previously the top bit of the size field was always 0, so the instruction classes only provided a 1-bit size field, which I have widened to 2 bits. Differential Revision: http://reviews.llvm.org/D15014 llvm-svn: 254198
This commit is contained in:
Oliver Stannard
parent
f2d3dbfaca
commit
b25914e03f
|
|
@ -352,9 +352,11 @@ class fixedpoint_i64<ValueType FloatVT>
|
|||
let ParserMatchClass = Imm1_64Operand;
|
||||
}
|
||||
|
||||
def fixedpoint_f16_i32 : fixedpoint_i32<f16>;
|
||||
def fixedpoint_f32_i32 : fixedpoint_i32<f32>;
|
||||
def fixedpoint_f64_i32 : fixedpoint_i32<f64>;
|
||||
|
||||
def fixedpoint_f16_i64 : fixedpoint_i64<f16>;
|
||||
def fixedpoint_f32_i64 : fixedpoint_i64<f32>;
|
||||
def fixedpoint_f64_i64 : fixedpoint_i64<f64>;
|
||||
|
||||
|
|
@ -713,6 +715,17 @@ class arith_extended_reg32to64<ValueType Ty> : Operand<Ty>,
|
|||
}
|
||||
|
||||
// Floating-point immediate.
|
||||
def fpimm16 : Operand<f16>,
|
||||
PatLeaf<(f16 fpimm), [{
|
||||
return AArch64_AM::getFP16Imm(N->getValueAPF()) != -1;
|
||||
}], SDNodeXForm<fpimm, [{
|
||||
APFloat InVal = N->getValueAPF();
|
||||
uint32_t enc = AArch64_AM::getFP16Imm(InVal);
|
||||
return CurDAG->getTargetConstant(enc, SDLoc(N), MVT::i32);
|
||||
}]>> {
|
||||
let ParserMatchClass = FPImmOperand;
|
||||
let PrintMethod = "printFPImmOperand";
|
||||
}
|
||||
def fpimm32 : Operand<f32>,
|
||||
PatLeaf<(f32 fpimm), [{
|
||||
return AArch64_AM::getFP32Imm(N->getValueAPF()) != -1;
|
||||
|
|
@ -3536,6 +3549,20 @@ class BaseFPToInteger<bits<2> type, bits<2> rmode, bits<3> opcode,
|
|||
|
||||
multiclass FPToIntegerUnscaled<bits<2> rmode, bits<3> opcode, string asm,
|
||||
SDPatternOperator OpN> {
|
||||
// Unscaled half-precision to 32-bit
|
||||
def UWHr : BaseFPToIntegerUnscaled<0b11, rmode, opcode, FPR16, GPR32, asm,
|
||||
[(set GPR32:$Rd, (OpN FPR16:$Rn))]> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
// Unscaled half-precision to 64-bit
|
||||
def UXHr : BaseFPToIntegerUnscaled<0b11, rmode, opcode, FPR16, GPR64, asm,
|
||||
[(set GPR64:$Rd, (OpN FPR16:$Rn))]> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
// Unscaled single-precision to 32-bit
|
||||
def UWSr : BaseFPToIntegerUnscaled<0b00, rmode, opcode, FPR32, GPR32, asm,
|
||||
[(set GPR32:$Rd, (OpN FPR32:$Rn))]> {
|
||||
|
|
@ -3563,6 +3590,25 @@ multiclass FPToIntegerUnscaled<bits<2> rmode, bits<3> opcode, string asm,
|
|||
|
||||
multiclass FPToIntegerScaled<bits<2> rmode, bits<3> opcode, string asm,
|
||||
SDPatternOperator OpN> {
|
||||
// Scaled half-precision to 32-bit
|
||||
def SWHri : BaseFPToInteger<0b11, rmode, opcode, FPR16, GPR32,
|
||||
fixedpoint_f16_i32, asm,
|
||||
[(set GPR32:$Rd, (OpN (fmul FPR16:$Rn,
|
||||
fixedpoint_f16_i32:$scale)))]> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let scale{5} = 1;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
// Scaled half-precision to 64-bit
|
||||
def SXHri : BaseFPToInteger<0b11, rmode, opcode, FPR16, GPR64,
|
||||
fixedpoint_f16_i64, asm,
|
||||
[(set GPR64:$Rd, (OpN (fmul FPR16:$Rn,
|
||||
fixedpoint_f16_i64:$scale)))]> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
// Scaled single-precision to 32-bit
|
||||
def SWSri : BaseFPToInteger<0b00, rmode, opcode, FPR32, GPR32,
|
||||
fixedpoint_f32_i32, asm,
|
||||
|
|
@ -3612,7 +3658,7 @@ class BaseIntegerToFP<bit isUnsigned,
|
|||
bits<5> Rd;
|
||||
bits<5> Rn;
|
||||
bits<6> scale;
|
||||
let Inst{30-23} = 0b00111100;
|
||||
let Inst{30-24} = 0b0011110;
|
||||
let Inst{21-17} = 0b00001;
|
||||
let Inst{16} = isUnsigned;
|
||||
let Inst{15-10} = scale;
|
||||
|
|
@ -3629,7 +3675,7 @@ class BaseIntegerToFPUnscaled<bit isUnsigned,
|
|||
bits<5> Rd;
|
||||
bits<5> Rn;
|
||||
bits<6> scale;
|
||||
let Inst{30-23} = 0b00111100;
|
||||
let Inst{30-24} = 0b0011110;
|
||||
let Inst{21-17} = 0b10001;
|
||||
let Inst{16} = isUnsigned;
|
||||
let Inst{15-10} = 0b000000;
|
||||
|
|
@ -3639,33 +3685,55 @@ class BaseIntegerToFPUnscaled<bit isUnsigned,
|
|||
|
||||
multiclass IntegerToFP<bit isUnsigned, string asm, SDNode node> {
|
||||
// Unscaled
|
||||
def UWHri: BaseIntegerToFPUnscaled<isUnsigned, GPR32, FPR16, f16, asm, node> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def UWSri: BaseIntegerToFPUnscaled<isUnsigned, GPR32, FPR32, f32, asm, node> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{22} = 0; // 32-bit FPR flag
|
||||
let Inst{23-22} = 0b00; // 32-bit FPR flag
|
||||
}
|
||||
|
||||
def UWDri: BaseIntegerToFPUnscaled<isUnsigned, GPR32, FPR64, f64, asm, node> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{22} = 1; // 64-bit FPR flag
|
||||
let Inst{23-22} = 0b01; // 64-bit FPR flag
|
||||
}
|
||||
|
||||
def UXHri: BaseIntegerToFPUnscaled<isUnsigned, GPR64, FPR16, f16, asm, node> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def UXSri: BaseIntegerToFPUnscaled<isUnsigned, GPR64, FPR32, f32, asm, node> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{22} = 0; // 32-bit FPR flag
|
||||
let Inst{23-22} = 0b00; // 32-bit FPR flag
|
||||
}
|
||||
|
||||
def UXDri: BaseIntegerToFPUnscaled<isUnsigned, GPR64, FPR64, f64, asm, node> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{22} = 1; // 64-bit FPR flag
|
||||
let Inst{23-22} = 0b01; // 64-bit FPR flag
|
||||
}
|
||||
|
||||
// Scaled
|
||||
def SWHri: BaseIntegerToFP<isUnsigned, GPR32, FPR16, fixedpoint_f16_i32, asm,
|
||||
[(set FPR16:$Rd,
|
||||
(fdiv (node GPR32:$Rn),
|
||||
fixedpoint_f16_i32:$scale))]> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let scale{5} = 1;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def SWSri: BaseIntegerToFP<isUnsigned, GPR32, FPR32, fixedpoint_f32_i32, asm,
|
||||
[(set FPR32:$Rd,
|
||||
(fdiv (node GPR32:$Rn),
|
||||
fixedpoint_f32_i32:$scale))]> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{22} = 0; // 32-bit FPR flag
|
||||
let Inst{23-22} = 0b00; // 32-bit FPR flag
|
||||
let scale{5} = 1;
|
||||
}
|
||||
|
||||
|
|
@ -3674,16 +3742,25 @@ multiclass IntegerToFP<bit isUnsigned, string asm, SDNode node> {
|
|||
(fdiv (node GPR32:$Rn),
|
||||
fixedpoint_f64_i32:$scale))]> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{22} = 1; // 64-bit FPR flag
|
||||
let Inst{23-22} = 0b01; // 64-bit FPR flag
|
||||
let scale{5} = 1;
|
||||
}
|
||||
|
||||
def SXHri: BaseIntegerToFP<isUnsigned, GPR64, FPR16, fixedpoint_f16_i64, asm,
|
||||
[(set FPR16:$Rd,
|
||||
(fdiv (node GPR64:$Rn),
|
||||
fixedpoint_f16_i64:$scale))]> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def SXSri: BaseIntegerToFP<isUnsigned, GPR64, FPR32, fixedpoint_f32_i64, asm,
|
||||
[(set FPR32:$Rd,
|
||||
(fdiv (node GPR64:$Rn),
|
||||
fixedpoint_f32_i64:$scale))]> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{22} = 0; // 32-bit FPR flag
|
||||
let Inst{23-22} = 0b00; // 32-bit FPR flag
|
||||
}
|
||||
|
||||
def SXDri: BaseIntegerToFP<isUnsigned, GPR64, FPR64, fixedpoint_f64_i64, asm,
|
||||
|
|
@ -3691,7 +3768,7 @@ multiclass IntegerToFP<bit isUnsigned, string asm, SDNode node> {
|
|||
(fdiv (node GPR64:$Rn),
|
||||
fixedpoint_f64_i64:$scale))]> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{22} = 1; // 64-bit FPR flag
|
||||
let Inst{23-22} = 0b01; // 64-bit FPR flag
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -3713,7 +3790,7 @@ class BaseUnscaledConversion<bits<2> rmode, bits<3> opcode,
|
|||
Sched<[WriteFCopy]> {
|
||||
bits<5> Rd;
|
||||
bits<5> Rn;
|
||||
let Inst{30-23} = 0b00111100;
|
||||
let Inst{30-24} = 0b0011110;
|
||||
let Inst{21} = 1;
|
||||
let Inst{20-19} = rmode;
|
||||
let Inst{18-16} = opcode;
|
||||
|
|
@ -3763,26 +3840,49 @@ class BaseUnscaledConversionFromHigh<bits<2> rmode, bits<3> opcode,
|
|||
}
|
||||
|
||||
|
||||
|
||||
multiclass UnscaledConversion<string asm> {
|
||||
def WHr : BaseUnscaledConversion<0b00, 0b111, GPR32, FPR16, asm> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def XHr : BaseUnscaledConversion<0b00, 0b111, GPR64, FPR16, asm> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def WSr : BaseUnscaledConversion<0b00, 0b111, GPR32, FPR32, asm> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{22} = 0; // 32-bit FPR flag
|
||||
let Inst{23-22} = 0b00; // 32-bit FPR flag
|
||||
}
|
||||
|
||||
def XDr : BaseUnscaledConversion<0b00, 0b111, GPR64, FPR64, asm> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{22} = 1; // 64-bit FPR flag
|
||||
let Inst{23-22} = 0b01; // 64-bit FPR flag
|
||||
}
|
||||
|
||||
def HWr : BaseUnscaledConversion<0b00, 0b110, FPR16, GPR32, asm> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def HXr : BaseUnscaledConversion<0b00, 0b110, FPR16, GPR64, asm> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{23-22} = 0b11; // 16-bit FPR flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def SWr : BaseUnscaledConversion<0b00, 0b110, FPR32, GPR32, asm> {
|
||||
let Inst{31} = 0; // 32-bit GPR flag
|
||||
let Inst{22} = 0; // 32-bit FPR flag
|
||||
let Inst{23-22} = 0b00; // 32-bit FPR flag
|
||||
}
|
||||
|
||||
def DXr : BaseUnscaledConversion<0b00, 0b110, FPR64, GPR64, asm> {
|
||||
let Inst{31} = 1; // 64-bit GPR flag
|
||||
let Inst{22} = 1; // 64-bit FPR flag
|
||||
let Inst{23-22} = 0b01; // 64-bit FPR flag
|
||||
}
|
||||
|
||||
def XDHighr : BaseUnscaledConversionToHigh<0b01, 0b111, GPR64, V128,
|
||||
|
|
@ -3855,7 +3955,7 @@ class BaseSingleOperandFPData<bits<4> opcode, RegisterClass regtype,
|
|||
Sched<[WriteF]> {
|
||||
bits<5> Rd;
|
||||
bits<5> Rn;
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21-19} = 0b100;
|
||||
let Inst{18-15} = opcode;
|
||||
let Inst{14-10} = 0b10000;
|
||||
|
|
@ -3865,12 +3965,17 @@ class BaseSingleOperandFPData<bits<4> opcode, RegisterClass regtype,
|
|||
|
||||
multiclass SingleOperandFPData<bits<4> opcode, string asm,
|
||||
SDPatternOperator node = null_frag> {
|
||||
def Hr : BaseSingleOperandFPData<opcode, FPR16, f16, asm, node> {
|
||||
let Inst{23-22} = 0b11; // 16-bit size flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Sr : BaseSingleOperandFPData<opcode, FPR32, f32, asm, node> {
|
||||
let Inst{22} = 0; // 32-bit size flag
|
||||
let Inst{23-22} = 0b00; // 32-bit size flag
|
||||
}
|
||||
|
||||
def Dr : BaseSingleOperandFPData<opcode, FPR64, f64, asm, node> {
|
||||
let Inst{22} = 1; // 64-bit size flag
|
||||
let Inst{23-22} = 0b01; // 64-bit size flag
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -3887,7 +3992,7 @@ class BaseTwoOperandFPData<bits<4> opcode, RegisterClass regtype,
|
|||
bits<5> Rd;
|
||||
bits<5> Rn;
|
||||
bits<5> Rm;
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21} = 1;
|
||||
let Inst{20-16} = Rm;
|
||||
let Inst{15-12} = opcode;
|
||||
|
|
@ -3898,28 +4003,41 @@ class BaseTwoOperandFPData<bits<4> opcode, RegisterClass regtype,
|
|||
|
||||
multiclass TwoOperandFPData<bits<4> opcode, string asm,
|
||||
SDPatternOperator node = null_frag> {
|
||||
def Hrr : BaseTwoOperandFPData<opcode, FPR16, asm,
|
||||
[(set (f16 FPR16:$Rd),
|
||||
(node (f16 FPR16:$Rn), (f16 FPR16:$Rm)))]> {
|
||||
let Inst{23-22} = 0b11; // 16-bit size flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Srr : BaseTwoOperandFPData<opcode, FPR32, asm,
|
||||
[(set (f32 FPR32:$Rd),
|
||||
(node (f32 FPR32:$Rn), (f32 FPR32:$Rm)))]> {
|
||||
let Inst{22} = 0; // 32-bit size flag
|
||||
let Inst{23-22} = 0b00; // 32-bit size flag
|
||||
}
|
||||
|
||||
def Drr : BaseTwoOperandFPData<opcode, FPR64, asm,
|
||||
[(set (f64 FPR64:$Rd),
|
||||
(node (f64 FPR64:$Rn), (f64 FPR64:$Rm)))]> {
|
||||
let Inst{22} = 1; // 64-bit size flag
|
||||
let Inst{23-22} = 0b01; // 64-bit size flag
|
||||
}
|
||||
}
|
||||
|
||||
multiclass TwoOperandFPDataNeg<bits<4> opcode, string asm, SDNode node> {
|
||||
def Hrr : BaseTwoOperandFPData<opcode, FPR16, asm,
|
||||
[(set FPR16:$Rd, (fneg (node FPR16:$Rn, (f16 FPR16:$Rm))))]> {
|
||||
let Inst{23-22} = 0b11; // 16-bit size flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Srr : BaseTwoOperandFPData<opcode, FPR32, asm,
|
||||
[(set FPR32:$Rd, (fneg (node FPR32:$Rn, (f32 FPR32:$Rm))))]> {
|
||||
let Inst{22} = 0; // 32-bit size flag
|
||||
let Inst{23-22} = 0b00; // 32-bit size flag
|
||||
}
|
||||
|
||||
def Drr : BaseTwoOperandFPData<opcode, FPR64, asm,
|
||||
[(set FPR64:$Rd, (fneg (node FPR64:$Rn, (f64 FPR64:$Rm))))]> {
|
||||
let Inst{22} = 1; // 64-bit size flag
|
||||
let Inst{23-22} = 0b01; // 64-bit size flag
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -3937,7 +4055,7 @@ class BaseThreeOperandFPData<bit isNegated, bit isSub,
|
|||
bits<5> Rn;
|
||||
bits<5> Rm;
|
||||
bits<5> Ra;
|
||||
let Inst{31-23} = 0b000111110;
|
||||
let Inst{31-24} = 0b00011111;
|
||||
let Inst{21} = isNegated;
|
||||
let Inst{20-16} = Rm;
|
||||
let Inst{15} = isSub;
|
||||
|
|
@ -3948,16 +4066,23 @@ class BaseThreeOperandFPData<bit isNegated, bit isSub,
|
|||
|
||||
multiclass ThreeOperandFPData<bit isNegated, bit isSub,string asm,
|
||||
SDPatternOperator node> {
|
||||
def Hrrr : BaseThreeOperandFPData<isNegated, isSub, FPR16, asm,
|
||||
[(set FPR16:$Rd,
|
||||
(node (f16 FPR16:$Rn), (f16 FPR16:$Rm), (f16 FPR16:$Ra)))]> {
|
||||
let Inst{23-22} = 0b11; // 16-bit size flag
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Srrr : BaseThreeOperandFPData<isNegated, isSub, FPR32, asm,
|
||||
[(set FPR32:$Rd,
|
||||
(node (f32 FPR32:$Rn), (f32 FPR32:$Rm), (f32 FPR32:$Ra)))]> {
|
||||
let Inst{22} = 0; // 32-bit size flag
|
||||
let Inst{23-22} = 0b00; // 32-bit size flag
|
||||
}
|
||||
|
||||
def Drrr : BaseThreeOperandFPData<isNegated, isSub, FPR64, asm,
|
||||
[(set FPR64:$Rd,
|
||||
(node (f64 FPR64:$Rn), (f64 FPR64:$Rm), (f64 FPR64:$Ra)))]> {
|
||||
let Inst{22} = 1; // 64-bit size flag
|
||||
let Inst{23-22} = 0b01; // 64-bit size flag
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -3972,7 +4097,7 @@ class BaseOneOperandFPComparison<bit signalAllNans,
|
|||
: I<(outs), (ins regtype:$Rn), asm, "\t$Rn, #0.0", "", pat>,
|
||||
Sched<[WriteFCmp]> {
|
||||
bits<5> Rn;
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21} = 1;
|
||||
|
||||
let Inst{15-10} = 0b001000;
|
||||
|
|
@ -3991,7 +4116,7 @@ class BaseTwoOperandFPComparison<bit signalAllNans, RegisterClass regtype,
|
|||
Sched<[WriteFCmp]> {
|
||||
bits<5> Rm;
|
||||
bits<5> Rn;
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21} = 1;
|
||||
let Inst{20-16} = Rm;
|
||||
let Inst{15-10} = 0b001000;
|
||||
|
|
@ -4003,24 +4128,36 @@ class BaseTwoOperandFPComparison<bit signalAllNans, RegisterClass regtype,
|
|||
multiclass FPComparison<bit signalAllNans, string asm,
|
||||
SDPatternOperator OpNode = null_frag> {
|
||||
let Defs = [NZCV] in {
|
||||
def Hrr : BaseTwoOperandFPComparison<signalAllNans, FPR16, asm,
|
||||
[(OpNode FPR16:$Rn, (f16 FPR16:$Rm)), (implicit NZCV)]> {
|
||||
let Inst{23-22} = 0b11;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Hri : BaseOneOperandFPComparison<signalAllNans, FPR16, asm,
|
||||
[(OpNode (f16 FPR16:$Rn), fpimm0), (implicit NZCV)]> {
|
||||
let Inst{23-22} = 0b11;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Srr : BaseTwoOperandFPComparison<signalAllNans, FPR32, asm,
|
||||
[(OpNode FPR32:$Rn, (f32 FPR32:$Rm)), (implicit NZCV)]> {
|
||||
let Inst{22} = 0;
|
||||
let Inst{23-22} = 0b00;
|
||||
}
|
||||
|
||||
def Sri : BaseOneOperandFPComparison<signalAllNans, FPR32, asm,
|
||||
[(OpNode (f32 FPR32:$Rn), fpimm0), (implicit NZCV)]> {
|
||||
let Inst{22} = 0;
|
||||
let Inst{23-22} = 0b00;
|
||||
}
|
||||
|
||||
def Drr : BaseTwoOperandFPComparison<signalAllNans, FPR64, asm,
|
||||
[(OpNode FPR64:$Rn, (f64 FPR64:$Rm)), (implicit NZCV)]> {
|
||||
let Inst{22} = 1;
|
||||
let Inst{23-22} = 0b01;
|
||||
}
|
||||
|
||||
def Dri : BaseOneOperandFPComparison<signalAllNans, FPR64, asm,
|
||||
[(OpNode (f64 FPR64:$Rn), fpimm0), (implicit NZCV)]> {
|
||||
let Inst{22} = 1;
|
||||
let Inst{23-22} = 0b01;
|
||||
}
|
||||
} // Defs = [NZCV]
|
||||
}
|
||||
|
|
@ -4043,7 +4180,7 @@ class BaseFPCondComparison<bit signalAllNans, RegisterClass regtype,
|
|||
bits<4> nzcv;
|
||||
bits<4> cond;
|
||||
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21} = 1;
|
||||
let Inst{20-16} = Rm;
|
||||
let Inst{15-12} = cond;
|
||||
|
|
@ -4055,15 +4192,21 @@ class BaseFPCondComparison<bit signalAllNans, RegisterClass regtype,
|
|||
|
||||
multiclass FPCondComparison<bit signalAllNans, string mnemonic,
|
||||
SDPatternOperator OpNode = null_frag> {
|
||||
def Hrr : BaseFPCondComparison<signalAllNans, FPR16, mnemonic, []> {
|
||||
let Inst{23-22} = 0b11;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Srr : BaseFPCondComparison<signalAllNans, FPR32, mnemonic,
|
||||
[(set NZCV, (OpNode (f32 FPR32:$Rn), (f32 FPR32:$Rm), (i32 imm:$nzcv),
|
||||
(i32 imm:$cond), NZCV))]> {
|
||||
let Inst{22} = 0;
|
||||
let Inst{23-22} = 0b00;
|
||||
}
|
||||
|
||||
def Drr : BaseFPCondComparison<signalAllNans, FPR64, mnemonic,
|
||||
[(set NZCV, (OpNode (f64 FPR64:$Rn), (f64 FPR64:$Rm), (i32 imm:$nzcv),
|
||||
(i32 imm:$cond), NZCV))]> {
|
||||
let Inst{22} = 1;
|
||||
let Inst{23-22} = 0b01;
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -4083,7 +4226,7 @@ class BaseFPCondSelect<RegisterClass regtype, ValueType vt, string asm>
|
|||
bits<5> Rm;
|
||||
bits<4> cond;
|
||||
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21} = 1;
|
||||
let Inst{20-16} = Rm;
|
||||
let Inst{15-12} = cond;
|
||||
|
|
@ -4094,12 +4237,17 @@ class BaseFPCondSelect<RegisterClass regtype, ValueType vt, string asm>
|
|||
|
||||
multiclass FPCondSelect<string asm> {
|
||||
let Uses = [NZCV] in {
|
||||
def Hrrr : BaseFPCondSelect<FPR16, f16, asm> {
|
||||
let Inst{23-22} = 0b11;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Srrr : BaseFPCondSelect<FPR32, f32, asm> {
|
||||
let Inst{22} = 0;
|
||||
let Inst{23-22} = 0b00;
|
||||
}
|
||||
|
||||
def Drrr : BaseFPCondSelect<FPR64, f64, asm> {
|
||||
let Inst{22} = 1;
|
||||
let Inst{23-22} = 0b01;
|
||||
}
|
||||
} // Uses = [NZCV]
|
||||
}
|
||||
|
|
@ -4114,7 +4262,7 @@ class BaseFPMoveImmediate<RegisterClass regtype, Operand fpimmtype, string asm>
|
|||
Sched<[WriteFImm]> {
|
||||
bits<5> Rd;
|
||||
bits<8> imm;
|
||||
let Inst{31-23} = 0b000111100;
|
||||
let Inst{31-24} = 0b00011110;
|
||||
let Inst{21} = 1;
|
||||
let Inst{20-13} = imm;
|
||||
let Inst{12-5} = 0b10000000;
|
||||
|
|
@ -4122,12 +4270,17 @@ class BaseFPMoveImmediate<RegisterClass regtype, Operand fpimmtype, string asm>
|
|||
}
|
||||
|
||||
multiclass FPMoveImmediate<string asm> {
|
||||
def Hi : BaseFPMoveImmediate<FPR16, fpimm16, asm> {
|
||||
let Inst{23-22} = 0b11;
|
||||
let Predicates = [HasFullFP16];
|
||||
}
|
||||
|
||||
def Si : BaseFPMoveImmediate<FPR32, fpimm32, asm> {
|
||||
let Inst{22} = 0;
|
||||
let Inst{23-22} = 0b00;
|
||||
}
|
||||
|
||||
def Di : BaseFPMoveImmediate<FPR64, fpimm64, asm> {
|
||||
let Inst{22} = 1;
|
||||
let Inst{23-22} = 0b01;
|
||||
}
|
||||
}
|
||||
} // end of 'let Predicates = [HasFPARMv8]'
|
||||
|
|
|
|||
|
|
@ -3963,7 +3963,7 @@ bool AArch64AsmParser::MatchAndEmitInstruction(SMLoc IDLoc, unsigned &Opcode,
|
|||
AArch64Operand &ImmOp = static_cast<AArch64Operand &>(*Operands[2]);
|
||||
if (RegOp.isReg() && ImmOp.isFPImm() && ImmOp.getFPImm() == (unsigned)-1) {
|
||||
unsigned zreg =
|
||||
AArch64MCRegisterClasses[AArch64::FPR32RegClassID].contains(
|
||||
!AArch64MCRegisterClasses[AArch64::FPR64RegClassID].contains(
|
||||
RegOp.getReg())
|
||||
? AArch64::WZR
|
||||
: AArch64::XZR;
|
||||
|
|
|
|||
|
|
@ -364,6 +364,32 @@ static inline float getFPImmFloat(unsigned Imm) {
|
|||
return FPUnion.F;
|
||||
}
|
||||
|
||||
/// getFP16Imm - Return an 8-bit floating-point version of the 16-bit
|
||||
/// floating-point value. If the value cannot be represented as an 8-bit
|
||||
/// floating-point value, then return -1.
|
||||
static inline int getFP16Imm(const APInt &Imm) {
|
||||
uint32_t Sign = Imm.lshr(15).getZExtValue() & 1;
|
||||
int32_t Exp = (Imm.lshr(10).getSExtValue() & 0x1f) - 15; // -14 to 15
|
||||
int32_t Mantissa = Imm.getZExtValue() & 0x3ff; // 10 bits
|
||||
|
||||
// We can handle 4 bits of mantissa.
|
||||
// mantissa = (16+UInt(e:f:g:h))/16.
|
||||
if (Mantissa & 0x3f)
|
||||
return -1;
|
||||
Mantissa >>= 6;
|
||||
|
||||
// We can handle 3 bits of exponent: exp == UInt(NOT(b):c:d)-3
|
||||
if (Exp < -3 || Exp > 4)
|
||||
return -1;
|
||||
Exp = ((Exp+3) & 0x7) ^ 4;
|
||||
|
||||
return ((int)Sign << 7) | (Exp << 4) | Mantissa;
|
||||
}
|
||||
|
||||
static inline int getFP16Imm(const APFloat &FPImm) {
|
||||
return getFP16Imm(FPImm.bitcastToAPInt());
|
||||
}
|
||||
|
||||
/// getFP32Imm - Return an 8-bit floating-point version of the 32-bit
|
||||
/// floating-point value. If the value cannot be represented as an 8-bit
|
||||
/// floating-point value, then return -1.
|
||||
|
|
|
|||
|
|
@ -1,99 +1,165 @@
|
|||
; RUN: llvm-mc -triple arm64-apple-darwin -mattr=neon -show-encoding -output-asm-variant=1 < %s | FileCheck %s
|
||||
; RUN: not llvm-mc -triple arm64-apple-darwin -mattr=neon -show-encoding -output-asm-variant=1 < %s 2>%t | FileCheck %s
|
||||
; RUN: FileCheck %s < %t --check-prefix=NO-FP16
|
||||
; RUN: llvm-mc -triple arm64-apple-darwin -mattr=neon,v8.2a,fullfp16 -show-encoding -output-asm-variant=1 < %s | FileCheck %s --check-prefix=CHECK --check-prefix=FP16
|
||||
|
||||
foo:
|
||||
;-----------------------------------------------------------------------------
|
||||
; Floating-point arithmetic
|
||||
;-----------------------------------------------------------------------------
|
||||
|
||||
fabs h1, h2
|
||||
fabs s1, s2
|
||||
fabs d1, d2
|
||||
|
||||
; FP16: fabs h1, h2 ; encoding: [0x41,0xc0,0xe0,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fabs h1, h2
|
||||
; CHECK: fabs s1, s2 ; encoding: [0x41,0xc0,0x20,0x1e]
|
||||
; CHECK: fabs d1, d2 ; encoding: [0x41,0xc0,0x60,0x1e]
|
||||
|
||||
fadd h1, h2, h3
|
||||
fadd s1, s2, s3
|
||||
fadd d1, d2, d3
|
||||
|
||||
; FP16: fadd h1, h2, h3 ; encoding: [0x41,0x28,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fadd h1, h2, h3
|
||||
; CHECK: fadd s1, s2, s3 ; encoding: [0x41,0x28,0x23,0x1e]
|
||||
; CHECK: fadd d1, d2, d3 ; encoding: [0x41,0x28,0x63,0x1e]
|
||||
|
||||
fdiv h1, h2, h3
|
||||
fdiv s1, s2, s3
|
||||
fdiv d1, d2, d3
|
||||
|
||||
; FP16: fdiv h1, h2, h3 ; encoding: [0x41,0x18,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fdiv h1, h2, h3
|
||||
; CHECK: fdiv s1, s2, s3 ; encoding: [0x41,0x18,0x23,0x1e]
|
||||
; CHECK: fdiv d1, d2, d3 ; encoding: [0x41,0x18,0x63,0x1e]
|
||||
|
||||
fmadd h1, h2, h3, h4
|
||||
fmadd s1, s2, s3, s4
|
||||
fmadd d1, d2, d3, d4
|
||||
|
||||
; FP16: fmadd h1, h2, h3, h4 ; encoding: [0x41,0x10,0xc3,0x1f]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmadd h1, h2, h3, h4
|
||||
; CHECK: fmadd s1, s2, s3, s4 ; encoding: [0x41,0x10,0x03,0x1f]
|
||||
; CHECK: fmadd d1, d2, d3, d4 ; encoding: [0x41,0x10,0x43,0x1f]
|
||||
|
||||
fmax h1, h2, h3
|
||||
fmax s1, s2, s3
|
||||
fmax d1, d2, d3
|
||||
fmaxnm h1, h2, h3
|
||||
fmaxnm s1, s2, s3
|
||||
fmaxnm d1, d2, d3
|
||||
|
||||
; FP16: fmax h1, h2, h3 ; encoding: [0x41,0x48,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmax h1, h2, h3
|
||||
; CHECK: fmax s1, s2, s3 ; encoding: [0x41,0x48,0x23,0x1e]
|
||||
; CHECK: fmax d1, d2, d3 ; encoding: [0x41,0x48,0x63,0x1e]
|
||||
; FP16: fmaxnm h1, h2, h3 ; encoding: [0x41,0x68,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmaxnm h1, h2, h3
|
||||
; CHECK: fmaxnm s1, s2, s3 ; encoding: [0x41,0x68,0x23,0x1e]
|
||||
; CHECK: fmaxnm d1, d2, d3 ; encoding: [0x41,0x68,0x63,0x1e]
|
||||
|
||||
fmin h1, h2, h3
|
||||
fmin s1, s2, s3
|
||||
fmin d1, d2, d3
|
||||
fminnm h1, h2, h3
|
||||
fminnm s1, s2, s3
|
||||
fminnm d1, d2, d3
|
||||
|
||||
; FP16: fmin h1, h2, h3 ; encoding: [0x41,0x58,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmin h1, h2, h3
|
||||
; CHECK: fmin s1, s2, s3 ; encoding: [0x41,0x58,0x23,0x1e]
|
||||
; CHECK: fmin d1, d2, d3 ; encoding: [0x41,0x58,0x63,0x1e]
|
||||
; FP16: fminnm h1, h2, h3 ; encoding: [0x41,0x78,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fminnm h1, h2, h3
|
||||
; CHECK: fminnm s1, s2, s3 ; encoding: [0x41,0x78,0x23,0x1e]
|
||||
; CHECK: fminnm d1, d2, d3 ; encoding: [0x41,0x78,0x63,0x1e]
|
||||
|
||||
fmsub h1, h2, h3, h4
|
||||
fmsub s1, s2, s3, s4
|
||||
fmsub d1, d2, d3, d4
|
||||
|
||||
; FP16: fmsub h1, h2, h3, h4 ; encoding: [0x41,0x90,0xc3,0x1f]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmsub h1, h2, h3, h4
|
||||
; CHECK: fmsub s1, s2, s3, s4 ; encoding: [0x41,0x90,0x03,0x1f]
|
||||
; CHECK: fmsub d1, d2, d3, d4 ; encoding: [0x41,0x90,0x43,0x1f]
|
||||
|
||||
fmul h1, h2, h3
|
||||
fmul s1, s2, s3
|
||||
fmul d1, d2, d3
|
||||
|
||||
; FP16: fmul h1, h2, h3 ; encoding: [0x41,0x08,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmul h1, h2, h3
|
||||
; CHECK: fmul s1, s2, s3 ; encoding: [0x41,0x08,0x23,0x1e]
|
||||
; CHECK: fmul d1, d2, d3 ; encoding: [0x41,0x08,0x63,0x1e]
|
||||
|
||||
fneg h1, h2
|
||||
fneg s1, s2
|
||||
fneg d1, d2
|
||||
|
||||
; FP16: fneg h1, h2 ; encoding: [0x41,0x40,0xe1,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fneg h1, h2
|
||||
; CHECK: fneg s1, s2 ; encoding: [0x41,0x40,0x21,0x1e]
|
||||
; CHECK: fneg d1, d2 ; encoding: [0x41,0x40,0x61,0x1e]
|
||||
|
||||
fnmadd h1, h2, h3, h4
|
||||
fnmadd s1, s2, s3, s4
|
||||
fnmadd d1, d2, d3, d4
|
||||
|
||||
; FP16: fnmadd h1, h2, h3, h4 ; encoding: [0x41,0x10,0xe3,0x1f]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fnmadd h1, h2, h3, h4
|
||||
; CHECK: fnmadd s1, s2, s3, s4 ; encoding: [0x41,0x10,0x23,0x1f]
|
||||
; CHECK: fnmadd d1, d2, d3, d4 ; encoding: [0x41,0x10,0x63,0x1f]
|
||||
|
||||
fnmsub h1, h2, h3, h4
|
||||
fnmsub s1, s2, s3, s4
|
||||
fnmsub d1, d2, d3, d4
|
||||
|
||||
; FP16: fnmsub h1, h2, h3, h4 ; encoding: [0x41,0x90,0xe3,0x1f]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fnmsub h1, h2, h3, h4
|
||||
; CHECK: fnmsub s1, s2, s3, s4 ; encoding: [0x41,0x90,0x23,0x1f]
|
||||
; CHECK: fnmsub d1, d2, d3, d4 ; encoding: [0x41,0x90,0x63,0x1f]
|
||||
|
||||
fnmul h1, h2, h3
|
||||
fnmul s1, s2, s3
|
||||
fnmul d1, d2, d3
|
||||
|
||||
; FP16: fnmul h1, h2, h3 ; encoding: [0x41,0x88,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fnmul h1, h2, h3
|
||||
; CHECK: fnmul s1, s2, s3 ; encoding: [0x41,0x88,0x23,0x1e]
|
||||
; CHECK: fnmul d1, d2, d3 ; encoding: [0x41,0x88,0x63,0x1e]
|
||||
|
||||
fsqrt h1, h2
|
||||
fsqrt s1, s2
|
||||
fsqrt d1, d2
|
||||
|
||||
; FP16: fsqrt h1, h2 ; encoding: [0x41,0xc0,0xe1,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fsqrt h1, h2
|
||||
; CHECK: fsqrt s1, s2 ; encoding: [0x41,0xc0,0x21,0x1e]
|
||||
; CHECK: fsqrt d1, d2 ; encoding: [0x41,0xc0,0x61,0x1e]
|
||||
|
||||
fsub h1, h2, h3
|
||||
fsub s1, s2, s3
|
||||
fsub d1, d2, d3
|
||||
|
||||
; FP16: fsub h1, h2, h3 ; encoding: [0x41,0x38,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fsub h1, h2, h3
|
||||
; CHECK: fsub s1, s2, s3 ; encoding: [0x41,0x38,0x23,0x1e]
|
||||
; CHECK: fsub d1, d2, d3 ; encoding: [0x41,0x38,0x63,0x1e]
|
||||
|
||||
|
|
@ -101,31 +167,55 @@ foo:
|
|||
; Floating-point comparison
|
||||
;-----------------------------------------------------------------------------
|
||||
|
||||
fccmp h1, h2, #0, eq
|
||||
fccmp s1, s2, #0, eq
|
||||
fccmp d1, d2, #0, eq
|
||||
fccmpe h1, h2, #0, eq
|
||||
fccmpe s1, s2, #0, eq
|
||||
fccmpe d1, d2, #0, eq
|
||||
|
||||
; FP16: fccmp h1, h2, #0, eq ; encoding: [0x20,0x04,0xe2,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fccmp h1, h2, #0, eq
|
||||
; CHECK: fccmp s1, s2, #0, eq ; encoding: [0x20,0x04,0x22,0x1e]
|
||||
; CHECK: fccmp d1, d2, #0, eq ; encoding: [0x20,0x04,0x62,0x1e]
|
||||
; FP16: fccmpe h1, h2, #0, eq ; encoding: [0x30,0x04,0xe2,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fccmpe h1, h2, #0, eq
|
||||
; CHECK: fccmpe s1, s2, #0, eq ; encoding: [0x30,0x04,0x22,0x1e]
|
||||
; CHECK: fccmpe d1, d2, #0, eq ; encoding: [0x30,0x04,0x62,0x1e]
|
||||
|
||||
fcmp h1, h2
|
||||
fcmp s1, s2
|
||||
fcmp d1, d2
|
||||
fcmp h1, #0.0
|
||||
fcmp s1, #0.0
|
||||
fcmp d1, #0.0
|
||||
fcmpe h1, h2
|
||||
fcmpe s1, s2
|
||||
fcmpe d1, d2
|
||||
fcmpe h1, #0.0
|
||||
fcmpe s1, #0.0
|
||||
fcmpe d1, #0.0
|
||||
|
||||
; FP16: fcmp h1, h2 ; encoding: [0x20,0x20,0xe2,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcmp h1, h2
|
||||
; CHECK: fcmp s1, s2 ; encoding: [0x20,0x20,0x22,0x1e]
|
||||
; CHECK: fcmp d1, d2 ; encoding: [0x20,0x20,0x62,0x1e]
|
||||
; FP16: fcmp h1, #0.0 ; encoding: [0x28,0x20,0xe0,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcmp h1, #0.0
|
||||
; CHECK: fcmp s1, #0.0 ; encoding: [0x28,0x20,0x20,0x1e]
|
||||
; CHECK: fcmp d1, #0.0 ; encoding: [0x28,0x20,0x60,0x1e]
|
||||
; FP16: fcmpe h1, h2 ; encoding: [0x30,0x20,0xe2,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcmpe h1, h2
|
||||
; CHECK: fcmpe s1, s2 ; encoding: [0x30,0x20,0x22,0x1e]
|
||||
; CHECK: fcmpe d1, d2 ; encoding: [0x30,0x20,0x62,0x1e]
|
||||
; FP16: fcmpe h1, #0.0 ; encoding: [0x38,0x20,0xe0,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcmpe h1, #0.0
|
||||
; CHECK: fcmpe s1, #0.0 ; encoding: [0x38,0x20,0x20,0x1e]
|
||||
; CHECK: fcmpe d1, #0.0 ; encoding: [0x38,0x20,0x60,0x1e]
|
||||
|
||||
|
|
@ -133,9 +223,13 @@ foo:
|
|||
; Floating-point conditional select
|
||||
;-----------------------------------------------------------------------------
|
||||
|
||||
fcsel h1, h2, h3, eq
|
||||
fcsel s1, s2, s3, eq
|
||||
fcsel d1, d2, d3, eq
|
||||
|
||||
; FP16: fcsel h1, h2, h3, eq ; encoding: [0x41,0x0c,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcsel h1, h2, h3, eq
|
||||
; CHECK: fcsel s1, s2, s3, eq ; encoding: [0x41,0x0c,0x23,0x1e]
|
||||
; CHECK: fcsel d1, d2, d3, eq ; encoding: [0x41,0x0c,0x63,0x1e]
|
||||
|
||||
|
|
@ -161,168 +255,314 @@ foo:
|
|||
fcvtas x1, d2
|
||||
fcvtas w1, s2
|
||||
fcvtas x1, s2
|
||||
fcvtas w1, h2
|
||||
fcvtas x1, h2
|
||||
|
||||
; CHECK: fcvtas w1, d2 ; encoding: [0x41,0x00,0x64,0x1e]
|
||||
; CHECK: fcvtas x1, d2 ; encoding: [0x41,0x00,0x64,0x9e]
|
||||
; CHECK: fcvtas w1, s2 ; encoding: [0x41,0x00,0x24,0x1e]
|
||||
; CHECK: fcvtas x1, s2 ; encoding: [0x41,0x00,0x24,0x9e]
|
||||
; CHECK: fcvtas w1, d2 ; encoding: [0x41,0x00,0x64,0x1e]
|
||||
; CHECK: fcvtas x1, d2 ; encoding: [0x41,0x00,0x64,0x9e]
|
||||
; CHECK: fcvtas w1, s2 ; encoding: [0x41,0x00,0x24,0x1e]
|
||||
; CHECK: fcvtas x1, s2 ; encoding: [0x41,0x00,0x24,0x9e]
|
||||
; FP16: fcvtas w1, h2 ; encoding: [0x41,0x00,0xe4,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtas w1, h2
|
||||
; FP16: fcvtas x1, h2 ; encoding: [0x41,0x00,0xe4,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtas x1, h2
|
||||
|
||||
fcvtau w1, h2
|
||||
fcvtau w1, s2
|
||||
fcvtau w1, d2
|
||||
fcvtau x1, h2
|
||||
fcvtau x1, s2
|
||||
fcvtau x1, d2
|
||||
|
||||
; CHECK: fcvtau w1, s2 ; encoding: [0x41,0x00,0x25,0x1e]
|
||||
; CHECK: fcvtau w1, d2 ; encoding: [0x41,0x00,0x65,0x1e]
|
||||
; CHECK: fcvtau x1, s2 ; encoding: [0x41,0x00,0x25,0x9e]
|
||||
; CHECK: fcvtau x1, d2 ; encoding: [0x41,0x00,0x65,0x9e]
|
||||
; FP16: fcvtau w1, h2 ; encoding: [0x41,0x00,0xe5,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtau w1, h2
|
||||
; CHECK: fcvtau w1, s2 ; encoding: [0x41,0x00,0x25,0x1e]
|
||||
; CHECK: fcvtau w1, d2 ; encoding: [0x41,0x00,0x65,0x1e]
|
||||
; FP16: fcvtau x1, h2 ; encoding: [0x41,0x00,0xe5,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtau x1, h2
|
||||
; CHECK: fcvtau x1, s2 ; encoding: [0x41,0x00,0x25,0x9e]
|
||||
; CHECK: fcvtau x1, d2 ; encoding: [0x41,0x00,0x65,0x9e]
|
||||
|
||||
fcvtms w1, h2
|
||||
fcvtms w1, s2
|
||||
fcvtms w1, d2
|
||||
fcvtms x1, h2
|
||||
fcvtms x1, s2
|
||||
fcvtms x1, d2
|
||||
|
||||
; CHECK: fcvtms w1, s2 ; encoding: [0x41,0x00,0x30,0x1e]
|
||||
; CHECK: fcvtms w1, d2 ; encoding: [0x41,0x00,0x70,0x1e]
|
||||
; CHECK: fcvtms x1, s2 ; encoding: [0x41,0x00,0x30,0x9e]
|
||||
; CHECK: fcvtms x1, d2 ; encoding: [0x41,0x00,0x70,0x9e]
|
||||
; FP16: fcvtms w1, h2 ; encoding: [0x41,0x00,0xf0,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtms w1, h2
|
||||
; CHECK: fcvtms w1, s2 ; encoding: [0x41,0x00,0x30,0x1e]
|
||||
; CHECK: fcvtms w1, d2 ; encoding: [0x41,0x00,0x70,0x1e]
|
||||
; FP16: fcvtms x1, h2 ; encoding: [0x41,0x00,0xf0,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtms x1, h2
|
||||
; CHECK: fcvtms x1, s2 ; encoding: [0x41,0x00,0x30,0x9e]
|
||||
; CHECK: fcvtms x1, d2 ; encoding: [0x41,0x00,0x70,0x9e]
|
||||
|
||||
fcvtmu w1, h2
|
||||
fcvtmu w1, s2
|
||||
fcvtmu w1, d2
|
||||
fcvtmu x1, h2
|
||||
fcvtmu x1, s2
|
||||
fcvtmu x1, d2
|
||||
|
||||
; CHECK: fcvtmu w1, s2 ; encoding: [0x41,0x00,0x31,0x1e]
|
||||
; CHECK: fcvtmu w1, d2 ; encoding: [0x41,0x00,0x71,0x1e]
|
||||
; CHECK: fcvtmu x1, s2 ; encoding: [0x41,0x00,0x31,0x9e]
|
||||
; CHECK: fcvtmu x1, d2 ; encoding: [0x41,0x00,0x71,0x9e]
|
||||
; FP16: fcvtmu w1, h2 ; encoding: [0x41,0x00,0xf1,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtmu w1, h2
|
||||
; CHECK: fcvtmu w1, s2 ; encoding: [0x41,0x00,0x31,0x1e]
|
||||
; CHECK: fcvtmu w1, d2 ; encoding: [0x41,0x00,0x71,0x1e]
|
||||
; FP16: fcvtmu x1, h2 ; encoding: [0x41,0x00,0xf1,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtmu x1, h2
|
||||
; CHECK: fcvtmu x1, s2 ; encoding: [0x41,0x00,0x31,0x9e]
|
||||
; CHECK: fcvtmu x1, d2 ; encoding: [0x41,0x00,0x71,0x9e]
|
||||
|
||||
fcvtns w1, h2
|
||||
fcvtns w1, s2
|
||||
fcvtns w1, d2
|
||||
fcvtns x1, h2
|
||||
fcvtns x1, s2
|
||||
fcvtns x1, d2
|
||||
|
||||
; CHECK: fcvtns w1, s2 ; encoding: [0x41,0x00,0x20,0x1e]
|
||||
; CHECK: fcvtns w1, d2 ; encoding: [0x41,0x00,0x60,0x1e]
|
||||
; CHECK: fcvtns x1, s2 ; encoding: [0x41,0x00,0x20,0x9e]
|
||||
; CHECK: fcvtns x1, d2 ; encoding: [0x41,0x00,0x60,0x9e]
|
||||
; FP16: fcvtns w1, h2 ; encoding: [0x41,0x00,0xe0,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtns w1, h2
|
||||
; CHECK: fcvtns w1, s2 ; encoding: [0x41,0x00,0x20,0x1e]
|
||||
; CHECK: fcvtns w1, d2 ; encoding: [0x41,0x00,0x60,0x1e]
|
||||
; FP16: fcvtns x1, h2 ; encoding: [0x41,0x00,0xe0,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtns x1, h2
|
||||
; CHECK: fcvtns x1, s2 ; encoding: [0x41,0x00,0x20,0x9e]
|
||||
; CHECK: fcvtns x1, d2 ; encoding: [0x41,0x00,0x60,0x9e]
|
||||
|
||||
fcvtnu w1, h2
|
||||
fcvtnu w1, s2
|
||||
fcvtnu w1, d2
|
||||
fcvtnu x1, h2
|
||||
fcvtnu x1, s2
|
||||
fcvtnu x1, d2
|
||||
|
||||
; CHECK: fcvtnu w1, s2 ; encoding: [0x41,0x00,0x21,0x1e]
|
||||
; CHECK: fcvtnu w1, d2 ; encoding: [0x41,0x00,0x61,0x1e]
|
||||
; CHECK: fcvtnu x1, s2 ; encoding: [0x41,0x00,0x21,0x9e]
|
||||
; CHECK: fcvtnu x1, d2 ; encoding: [0x41,0x00,0x61,0x9e]
|
||||
; FP16: fcvtnu w1, h2 ; encoding: [0x41,0x00,0xe1,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtnu w1, h2
|
||||
; CHECK: fcvtnu w1, s2 ; encoding: [0x41,0x00,0x21,0x1e]
|
||||
; CHECK: fcvtnu w1, d2 ; encoding: [0x41,0x00,0x61,0x1e]
|
||||
; FP16: fcvtnu x1, h2 ; encoding: [0x41,0x00,0xe1,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtnu x1, h2
|
||||
; CHECK: fcvtnu x1, s2 ; encoding: [0x41,0x00,0x21,0x9e]
|
||||
; CHECK: fcvtnu x1, d2 ; encoding: [0x41,0x00,0x61,0x9e]
|
||||
|
||||
fcvtps w1, h2
|
||||
fcvtps w1, s2
|
||||
fcvtps w1, d2
|
||||
fcvtps x1, h2
|
||||
fcvtps x1, s2
|
||||
fcvtps x1, d2
|
||||
|
||||
; CHECK: fcvtps w1, s2 ; encoding: [0x41,0x00,0x28,0x1e]
|
||||
; CHECK: fcvtps w1, d2 ; encoding: [0x41,0x00,0x68,0x1e]
|
||||
; CHECK: fcvtps x1, s2 ; encoding: [0x41,0x00,0x28,0x9e]
|
||||
; CHECK: fcvtps x1, d2 ; encoding: [0x41,0x00,0x68,0x9e]
|
||||
; FP16: fcvtps w1, h2 ; encoding: [0x41,0x00,0xe8,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtps w1, h2
|
||||
; CHECK: fcvtps w1, s2 ; encoding: [0x41,0x00,0x28,0x1e]
|
||||
; CHECK: fcvtps w1, d2 ; encoding: [0x41,0x00,0x68,0x1e]
|
||||
; FP16: fcvtps x1, h2 ; encoding: [0x41,0x00,0xe8,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtps x1, h2
|
||||
; CHECK: fcvtps x1, s2 ; encoding: [0x41,0x00,0x28,0x9e]
|
||||
; CHECK: fcvtps x1, d2 ; encoding: [0x41,0x00,0x68,0x9e]
|
||||
|
||||
fcvtpu w1, h2
|
||||
fcvtpu w1, s2
|
||||
fcvtpu w1, d2
|
||||
fcvtpu x1, h2
|
||||
fcvtpu x1, s2
|
||||
fcvtpu x1, d2
|
||||
|
||||
; CHECK: fcvtpu w1, s2 ; encoding: [0x41,0x00,0x29,0x1e]
|
||||
; CHECK: fcvtpu w1, d2 ; encoding: [0x41,0x00,0x69,0x1e]
|
||||
; CHECK: fcvtpu x1, s2 ; encoding: [0x41,0x00,0x29,0x9e]
|
||||
; CHECK: fcvtpu x1, d2 ; encoding: [0x41,0x00,0x69,0x9e]
|
||||
; FP16: fcvtpu w1, h2 ; encoding: [0x41,0x00,0xe9,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtpu w1, h2
|
||||
; CHECK: fcvtpu w1, s2 ; encoding: [0x41,0x00,0x29,0x1e]
|
||||
; CHECK: fcvtpu w1, d2 ; encoding: [0x41,0x00,0x69,0x1e]
|
||||
; FP16: fcvtpu x1, h2 ; encoding: [0x41,0x00,0xe9,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtpu x1, h2
|
||||
; CHECK: fcvtpu x1, s2 ; encoding: [0x41,0x00,0x29,0x9e]
|
||||
; CHECK: fcvtpu x1, d2 ; encoding: [0x41,0x00,0x69,0x9e]
|
||||
|
||||
fcvtzs w1, h2
|
||||
fcvtzs w1, h2, #1
|
||||
fcvtzs w1, s2
|
||||
fcvtzs w1, s2, #1
|
||||
fcvtzs w1, d2
|
||||
fcvtzs w1, d2, #1
|
||||
fcvtzs x1, h2
|
||||
fcvtzs x1, h2, #1
|
||||
fcvtzs x1, s2
|
||||
fcvtzs x1, s2, #1
|
||||
fcvtzs x1, d2
|
||||
fcvtzs x1, d2, #1
|
||||
|
||||
; CHECK: fcvtzs w1, s2 ; encoding: [0x41,0x00,0x38,0x1e]
|
||||
; CHECK: fcvtzs w1, s2, #1 ; encoding: [0x41,0xfc,0x18,0x1e]
|
||||
; CHECK: fcvtzs w1, d2 ; encoding: [0x41,0x00,0x78,0x1e]
|
||||
; CHECK: fcvtzs w1, d2, #1 ; encoding: [0x41,0xfc,0x58,0x1e]
|
||||
; CHECK: fcvtzs x1, s2 ; encoding: [0x41,0x00,0x38,0x9e]
|
||||
; CHECK: fcvtzs x1, s2, #1 ; encoding: [0x41,0xfc,0x18,0x9e]
|
||||
; CHECK: fcvtzs x1, d2 ; encoding: [0x41,0x00,0x78,0x9e]
|
||||
; CHECK: fcvtzs x1, d2, #1 ; encoding: [0x41,0xfc,0x58,0x9e]
|
||||
; FP16: fcvtzs w1, h2 ; encoding: [0x41,0x00,0xf8,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzs w1, h2
|
||||
; FP16: fcvtzs w1, h2, #1 ; encoding: [0x41,0xfc,0xd8,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzs w1, h2, #1
|
||||
; CHECK: fcvtzs w1, s2 ; encoding: [0x41,0x00,0x38,0x1e]
|
||||
; CHECK: fcvtzs w1, s2, #1 ; encoding: [0x41,0xfc,0x18,0x1e]
|
||||
; CHECK: fcvtzs w1, d2 ; encoding: [0x41,0x00,0x78,0x1e]
|
||||
; CHECK: fcvtzs w1, d2, #1 ; encoding: [0x41,0xfc,0x58,0x1e]
|
||||
; FP16: fcvtzs x1, h2 ; encoding: [0x41,0x00,0xf8,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzs x1, h2
|
||||
; FP16: fcvtzs x1, h2, #1 ; encoding: [0x41,0xfc,0xd8,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzs x1, h2, #1
|
||||
; CHECK: fcvtzs x1, s2 ; encoding: [0x41,0x00,0x38,0x9e]
|
||||
; CHECK: fcvtzs x1, s2, #1 ; encoding: [0x41,0xfc,0x18,0x9e]
|
||||
; CHECK: fcvtzs x1, d2 ; encoding: [0x41,0x00,0x78,0x9e]
|
||||
; CHECK: fcvtzs x1, d2, #1 ; encoding: [0x41,0xfc,0x58,0x9e]
|
||||
|
||||
fcvtzu w1, h2
|
||||
fcvtzu w1, h2, #1
|
||||
fcvtzu w1, s2
|
||||
fcvtzu w1, s2, #1
|
||||
fcvtzu w1, d2
|
||||
fcvtzu w1, d2, #1
|
||||
fcvtzu x1, h2
|
||||
fcvtzu x1, h2, #1
|
||||
fcvtzu x1, s2
|
||||
fcvtzu x1, s2, #1
|
||||
fcvtzu x1, d2
|
||||
fcvtzu x1, d2, #1
|
||||
|
||||
; CHECK: fcvtzu w1, s2 ; encoding: [0x41,0x00,0x39,0x1e]
|
||||
; CHECK: fcvtzu w1, s2, #1 ; encoding: [0x41,0xfc,0x19,0x1e]
|
||||
; CHECK: fcvtzu w1, d2 ; encoding: [0x41,0x00,0x79,0x1e]
|
||||
; CHECK: fcvtzu w1, d2, #1 ; encoding: [0x41,0xfc,0x59,0x1e]
|
||||
; CHECK: fcvtzu x1, s2 ; encoding: [0x41,0x00,0x39,0x9e]
|
||||
; CHECK: fcvtzu x1, s2, #1 ; encoding: [0x41,0xfc,0x19,0x9e]
|
||||
; CHECK: fcvtzu x1, d2 ; encoding: [0x41,0x00,0x79,0x9e]
|
||||
; CHECK: fcvtzu x1, d2, #1 ; encoding: [0x41,0xfc,0x59,0x9e]
|
||||
; FP16: fcvtzu w1, h2 ; encoding: [0x41,0x00,0xf9,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzu w1, h2
|
||||
; FP16: fcvtzu w1, h2, #1 ; encoding: [0x41,0xfc,0xd9,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzu w1, h2, #1
|
||||
; CHECK: fcvtzu w1, s2 ; encoding: [0x41,0x00,0x39,0x1e]
|
||||
; CHECK: fcvtzu w1, s2, #1 ; encoding: [0x41,0xfc,0x19,0x1e]
|
||||
; CHECK: fcvtzu w1, d2 ; encoding: [0x41,0x00,0x79,0x1e]
|
||||
; CHECK: fcvtzu w1, d2, #1 ; encoding: [0x41,0xfc,0x59,0x1e]
|
||||
; FP16: fcvtzu x1, h2 ; encoding: [0x41,0x00,0xf9,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzu x1, h2
|
||||
; FP16: fcvtzu x1, h2, #1 ; encoding: [0x41,0xfc,0xd9,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fcvtzu x1, h2, #1
|
||||
; CHECK: fcvtzu x1, s2 ; encoding: [0x41,0x00,0x39,0x9e]
|
||||
; CHECK: fcvtzu x1, s2, #1 ; encoding: [0x41,0xfc,0x19,0x9e]
|
||||
; CHECK: fcvtzu x1, d2 ; encoding: [0x41,0x00,0x79,0x9e]
|
||||
; CHECK: fcvtzu x1, d2, #1 ; encoding: [0x41,0xfc,0x59,0x9e]
|
||||
|
||||
scvtf h1, w2
|
||||
scvtf h1, w2, #1
|
||||
scvtf s1, w2
|
||||
scvtf s1, w2, #1
|
||||
scvtf d1, w2
|
||||
scvtf d1, w2, #1
|
||||
scvtf h1, x2
|
||||
scvtf h1, x2, #1
|
||||
scvtf s1, x2
|
||||
scvtf s1, x2, #1
|
||||
scvtf d1, x2
|
||||
scvtf d1, x2, #1
|
||||
|
||||
; CHECK: scvtf s1, w2 ; encoding: [0x41,0x00,0x22,0x1e]
|
||||
; CHECK: scvtf s1, w2, #1 ; encoding: [0x41,0xfc,0x02,0x1e]
|
||||
; CHECK: scvtf d1, w2 ; encoding: [0x41,0x00,0x62,0x1e]
|
||||
; CHECK: scvtf d1, w2, #1 ; encoding: [0x41,0xfc,0x42,0x1e]
|
||||
; CHECK: scvtf s1, x2 ; encoding: [0x41,0x00,0x22,0x9e]
|
||||
; CHECK: scvtf s1, x2, #1 ; encoding: [0x41,0xfc,0x02,0x9e]
|
||||
; CHECK: scvtf d1, x2 ; encoding: [0x41,0x00,0x62,0x9e]
|
||||
; CHECK: scvtf d1, x2, #1 ; encoding: [0x41,0xfc,0x42,0x9e]
|
||||
; FP16: scvtf h1, w2 ; encoding: [0x41,0x00,0xe2,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: scvtf h1, w2
|
||||
; FP16: scvtf h1, w2, #1 ; encoding: [0x41,0xfc,0xc2,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: scvtf h1, w2, #1
|
||||
; CHECK: scvtf s1, w2 ; encoding: [0x41,0x00,0x22,0x1e]
|
||||
; CHECK: scvtf s1, w2, #1 ; encoding: [0x41,0xfc,0x02,0x1e]
|
||||
; CHECK: scvtf d1, w2 ; encoding: [0x41,0x00,0x62,0x1e]
|
||||
; CHECK: scvtf d1, w2, #1 ; encoding: [0x41,0xfc,0x42,0x1e]
|
||||
; FP16: scvtf h1, x2 ; encoding: [0x41,0x00,0xe2,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: scvtf h1, x2
|
||||
; FP16: scvtf h1, x2, #1 ; encoding: [0x41,0xfc,0xc2,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: scvtf h1, x2, #1
|
||||
; CHECK: scvtf s1, x2 ; encoding: [0x41,0x00,0x22,0x9e]
|
||||
; CHECK: scvtf s1, x2, #1 ; encoding: [0x41,0xfc,0x02,0x9e]
|
||||
; CHECK: scvtf d1, x2 ; encoding: [0x41,0x00,0x62,0x9e]
|
||||
; CHECK: scvtf d1, x2, #1 ; encoding: [0x41,0xfc,0x42,0x9e]
|
||||
|
||||
ucvtf h1, w2
|
||||
ucvtf h1, w2, #1
|
||||
ucvtf s1, w2
|
||||
ucvtf s1, w2, #1
|
||||
ucvtf d1, w2
|
||||
ucvtf d1, w2, #1
|
||||
ucvtf h1, x2
|
||||
ucvtf h1, x2, #1
|
||||
ucvtf s1, x2
|
||||
ucvtf s1, x2, #1
|
||||
ucvtf d1, x2
|
||||
ucvtf d1, x2, #1
|
||||
|
||||
; CHECK: ucvtf s1, w2 ; encoding: [0x41,0x00,0x23,0x1e]
|
||||
; CHECK: ucvtf s1, w2, #1 ; encoding: [0x41,0xfc,0x03,0x1e]
|
||||
; CHECK: ucvtf d1, w2 ; encoding: [0x41,0x00,0x63,0x1e]
|
||||
; CHECK: ucvtf d1, w2, #1 ; encoding: [0x41,0xfc,0x43,0x1e]
|
||||
; CHECK: ucvtf s1, x2 ; encoding: [0x41,0x00,0x23,0x9e]
|
||||
; CHECK: ucvtf s1, x2, #1 ; encoding: [0x41,0xfc,0x03,0x9e]
|
||||
; CHECK: ucvtf d1, x2 ; encoding: [0x41,0x00,0x63,0x9e]
|
||||
; CHECK: ucvtf d1, x2, #1 ; encoding: [0x41,0xfc,0x43,0x9e]
|
||||
; FP16: ucvtf h1, w2 ; encoding: [0x41,0x00,0xe3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: ucvtf h1, w2
|
||||
; FP16: ucvtf h1, w2, #1 ; encoding: [0x41,0xfc,0xc3,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: ucvtf h1, w2, #1
|
||||
; CHECK: ucvtf s1, w2 ; encoding: [0x41,0x00,0x23,0x1e]
|
||||
; CHECK: ucvtf s1, w2, #1 ; encoding: [0x41,0xfc,0x03,0x1e]
|
||||
; CHECK: ucvtf d1, w2 ; encoding: [0x41,0x00,0x63,0x1e]
|
||||
; CHECK: ucvtf d1, w2, #1 ; encoding: [0x41,0xfc,0x43,0x1e]
|
||||
; FP16: ucvtf h1, x2 ; encoding: [0x41,0x00,0xe3,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: ucvtf h1, x2
|
||||
; FP16: ucvtf h1, x2, #1 ; encoding: [0x41,0xfc,0xc3,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: ucvtf h1, x2, #1
|
||||
; CHECK: ucvtf s1, x2 ; encoding: [0x41,0x00,0x23,0x9e]
|
||||
; CHECK: ucvtf s1, x2, #1 ; encoding: [0x41,0xfc,0x03,0x9e]
|
||||
; CHECK: ucvtf d1, x2 ; encoding: [0x41,0x00,0x63,0x9e]
|
||||
; CHECK: ucvtf d1, x2, #1 ; encoding: [0x41,0xfc,0x43,0x9e]
|
||||
|
||||
;-----------------------------------------------------------------------------
|
||||
; Floating-point move
|
||||
;-----------------------------------------------------------------------------
|
||||
|
||||
fmov h1, w2
|
||||
fmov w1, h2
|
||||
fmov h1, x2
|
||||
fmov x1, h2
|
||||
fmov s1, w2
|
||||
fmov w1, s2
|
||||
fmov d1, x2
|
||||
fmov x1, d2
|
||||
|
||||
; FP16: fmov h1, w2 ; encoding: [0x41,0x00,0xe7,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov h1, w2
|
||||
; FP16: fmov w1, h2 ; encoding: [0x41,0x00,0xe6,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov w1, h2
|
||||
; FP16: fmov h1, x2 ; encoding: [0x41,0x00,0xe7,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov h1, x2
|
||||
; FP16: fmov x1, h2 ; encoding: [0x41,0x00,0xe6,0x9e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov x1, h2
|
||||
; CHECK: fmov s1, w2 ; encoding: [0x41,0x00,0x27,0x1e]
|
||||
; CHECK: fmov w1, s2 ; encoding: [0x41,0x00,0x26,0x1e]
|
||||
; CHECK: fmov d1, x2 ; encoding: [0x41,0x00,0x67,0x9e]
|
||||
; CHECK: fmov x1, d2 ; encoding: [0x41,0x00,0x66,0x9e]
|
||||
|
||||
fmov h1, #0.125
|
||||
fmov h1, #0x40
|
||||
fmov s1, #0.125
|
||||
fmov s1, #0x40
|
||||
fmov d1, #0.125
|
||||
|
|
@ -330,9 +570,16 @@ foo:
|
|||
fmov d1, #-4.843750e-01
|
||||
fmov d1, #4.843750e-01
|
||||
fmov d3, #3
|
||||
fmov h2, #0.0
|
||||
fmov s2, #0.0
|
||||
fmov d2, #0.0
|
||||
|
||||
; FP16: fmov h1, #0.12500000 ; encoding: [0x01,0x10,0xe8,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov h1, #0.125
|
||||
; FP16: fmov h1, #0.12500000 ; encoding: [0x01,0x10,0xe8,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov h1, #0x40
|
||||
; CHECK: fmov s1, #0.12500000 ; encoding: [0x01,0x10,0x28,0x1e]
|
||||
; CHECK: fmov s1, #0.12500000 ; encoding: [0x01,0x10,0x28,0x1e]
|
||||
; CHECK: fmov d1, #0.12500000 ; encoding: [0x01,0x10,0x68,0x1e]
|
||||
|
|
@ -340,12 +587,19 @@ foo:
|
|||
; CHECK: fmov d1, #-0.48437500 ; encoding: [0x01,0xf0,0x7b,0x1e]
|
||||
; CHECK: fmov d1, #0.48437500 ; encoding: [0x01,0xf0,0x6b,0x1e]
|
||||
; CHECK: fmov d3, #3.00000000 ; encoding: [0x03,0x10,0x61,0x1e]
|
||||
; FP16: fmov h2, wzr ; encoding: [0xe2,0x03,0xe7,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov h2, #0.0
|
||||
; CHECK: fmov s2, wzr ; encoding: [0xe2,0x03,0x27,0x1e]
|
||||
; CHECK: fmov d2, xzr ; encoding: [0xe2,0x03,0x67,0x9e]
|
||||
|
||||
fmov h1, h2
|
||||
fmov s1, s2
|
||||
fmov d1, d2
|
||||
|
||||
; FP16: fmov h1, h2 ; encoding: [0x41,0x40,0xe0,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: fmov h1, h2
|
||||
; CHECK: fmov s1, s2 ; encoding: [0x41,0x40,0x20,0x1e]
|
||||
; CHECK: fmov d1, d2 ; encoding: [0x41,0x40,0x60,0x1e]
|
||||
|
||||
|
|
@ -355,63 +609,91 @@ foo:
|
|||
fmov v1.d[1], x1
|
||||
fmov.d v8[1], x6
|
||||
|
||||
; CHECK: fmov.d x2, v5[1] ; encoding: [0xa2,0x00,0xae,0x9e]
|
||||
; CHECK: fmov.d x9, v7[1] ; encoding: [0xe9,0x00,0xae,0x9e]
|
||||
; CHECK: fmov.d v1[1], x1 ; encoding: [0x21,0x00,0xaf,0x9e]
|
||||
; CHECK: fmov.d v8[1], x6 ; encoding: [0xc8,0x00,0xaf,0x9e]
|
||||
; CHECK: fmov.d x2, v5[1] ; encoding: [0xa2,0x00,0xae,0x9e]
|
||||
; CHECK: fmov.d x9, v7[1] ; encoding: [0xe9,0x00,0xae,0x9e]
|
||||
; CHECK: fmov.d v1[1], x1 ; encoding: [0x21,0x00,0xaf,0x9e]
|
||||
; CHECK: fmov.d v8[1], x6 ; encoding: [0xc8,0x00,0xaf,0x9e]
|
||||
|
||||
|
||||
;-----------------------------------------------------------------------------
|
||||
; Floating-point round to integral
|
||||
;-----------------------------------------------------------------------------
|
||||
|
||||
frinta h1, h2
|
||||
frinta s1, s2
|
||||
frinta d1, d2
|
||||
|
||||
; FP16: frinta h1, h2 ; encoding: [0x41,0x40,0xe6,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frinta h1, h2
|
||||
; CHECK: frinta s1, s2 ; encoding: [0x41,0x40,0x26,0x1e]
|
||||
; CHECK: frinta d1, d2 ; encoding: [0x41,0x40,0x66,0x1e]
|
||||
|
||||
frinti h1, h2
|
||||
frinti s1, s2
|
||||
frinti d1, d2
|
||||
|
||||
; FP16: frinti h1, h2 ; encoding: [0x41,0xc0,0xe7,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frinti h1, h2
|
||||
; CHECK: frinti s1, s2 ; encoding: [0x41,0xc0,0x27,0x1e]
|
||||
; CHECK: frinti d1, d2 ; encoding: [0x41,0xc0,0x67,0x1e]
|
||||
|
||||
frintm h1, h2
|
||||
frintm s1, s2
|
||||
frintm d1, d2
|
||||
|
||||
; FP16: frintm h1, h2 ; encoding: [0x41,0x40,0xe5,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frintm h1, h2
|
||||
; CHECK: frintm s1, s2 ; encoding: [0x41,0x40,0x25,0x1e]
|
||||
; CHECK: frintm d1, d2 ; encoding: [0x41,0x40,0x65,0x1e]
|
||||
|
||||
frintn h1, h2
|
||||
frintn s1, s2
|
||||
frintn d1, d2
|
||||
|
||||
; FP16: frintn h1, h2 ; encoding: [0x41,0x40,0xe4,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frintn h1, h2
|
||||
; CHECK: frintn s1, s2 ; encoding: [0x41,0x40,0x24,0x1e]
|
||||
; CHECK: frintn d1, d2 ; encoding: [0x41,0x40,0x64,0x1e]
|
||||
|
||||
frintp h1, h2
|
||||
frintp s1, s2
|
||||
frintp d1, d2
|
||||
|
||||
; FP16: frintp h1, h2 ; encoding: [0x41,0xc0,0xe4,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frintp h1, h2
|
||||
; CHECK: frintp s1, s2 ; encoding: [0x41,0xc0,0x24,0x1e]
|
||||
; CHECK: frintp d1, d2 ; encoding: [0x41,0xc0,0x64,0x1e]
|
||||
|
||||
frintx h1, h2
|
||||
frintx s1, s2
|
||||
frintx d1, d2
|
||||
|
||||
; FP16: frintx h1, h2 ; encoding: [0x41,0x40,0xe7,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frintx h1, h2
|
||||
; CHECK: frintx s1, s2 ; encoding: [0x41,0x40,0x27,0x1e]
|
||||
; CHECK: frintx d1, d2 ; encoding: [0x41,0x40,0x67,0x1e]
|
||||
|
||||
frintz h1, h2
|
||||
frintz s1, s2
|
||||
frintz d1, d2
|
||||
|
||||
; FP16: frintz h1, h2 ; encoding: [0x41,0xc0,0xe5,0x1e]
|
||||
; NO-FP16: error: instruction requires:
|
||||
; NO-FP16-NEXT: frintz h1, h2
|
||||
; CHECK: frintz s1, s2 ; encoding: [0x41,0xc0,0x25,0x1e]
|
||||
; CHECK: frintz d1, d2 ; encoding: [0x41,0xc0,0x65,0x1e]
|
||||
|
||||
cmhs d0, d0, d0
|
||||
cmtst d0, d0, d0
|
||||
|
||||
; CHECK: cmhs d0, d0, d0 ; encoding: [0x00,0x3c,0xe0,0x7e]
|
||||
; CHECK: cmtst d0, d0, d0 ; encoding: [0x00,0x8c,0xe0,0x5e]
|
||||
; CHECK: cmhs d0, d0, d0 ; encoding: [0x00,0x3c,0xe0,0x7e]
|
||||
; CHECK: cmtst d0, d0, d0 ; encoding: [0x00,0x8c,0xe0,0x5e]
|
||||
|
||||
|
||||
|
||||
|
|
|
|||
|
|
@ -1603,7 +1603,7 @@
|
|||
// CHECK-ERROR: error: invalid operand for instruction
|
||||
// CHECK-ERROR-NEXT: fcsel q3, q20, q9, pl
|
||||
// CHECK-ERROR-NEXT: ^
|
||||
// CHECK-ERROR-NEXT: error: invalid operand for instruction
|
||||
// CHECK-ERROR-NEXT: error: instruction requires: fullfp16
|
||||
// CHECK-ERROR-NEXT: fcsel h9, h10, h11, mi
|
||||
// CHECK-ERROR-NEXT: ^
|
||||
// CHECK-ERROR-NEXT: error: invalid operand for instruction
|
||||
|
|
@ -1652,7 +1652,7 @@
|
|||
// CHECK-ERROR: error: invalid operand for instruction
|
||||
// CHECK-ERROR-NEXT: fmadd b3, b4, b5, b6
|
||||
// CHECK-ERROR-NEXT: ^
|
||||
// CHECK-ERROR-NEXT: error: invalid operand for instruction
|
||||
// CHECK-ERROR-NEXT: error: instruction requires: fullfp16
|
||||
// CHECK-ERROR-NEXT: fmsub h1, h2, h3, h4
|
||||
// CHECK-ERROR-NEXT: ^
|
||||
// CHECK-ERROR-NEXT: error: invalid operand for instruction
|
||||
|
|
|
|||
|
|
@ -1,98 +1,131 @@
|
|||
# RUN: llvm-mc -triple arm64-apple-darwin -mattr=neon --disassemble -output-asm-variant=1 < %s | FileCheck %s
|
||||
# RUN: llvm-mc -triple arm64-apple-darwin -mattr=neon,v8.2a,fullfp16 --disassemble -output-asm-variant=1 < %s | FileCheck %s --check-prefix=CHECK --check-prefix=FP16
|
||||
|
||||
#-----------------------------------------------------------------------------
|
||||
# Floating-point arithmetic
|
||||
#-----------------------------------------------------------------------------
|
||||
|
||||
0x41 0xc0 0xe0 0x1e
|
||||
0x41 0xc0 0x20 0x1e
|
||||
0x41 0xc0 0x60 0x1e
|
||||
|
||||
# FP16: fabs h1, h2
|
||||
# CHECK: fabs s1, s2
|
||||
# CHECK: fabs d1, d2
|
||||
|
||||
0x41 0x28 0xe3 0x1e
|
||||
0x41 0x28 0x23 0x1e
|
||||
0x41 0x28 0x63 0x1e
|
||||
|
||||
# FP16: fadd h1, h2, h3
|
||||
# CHECK: fadd s1, s2, s3
|
||||
# CHECK: fadd d1, d2, d3
|
||||
|
||||
0x41 0x18 0xe3 0x1e
|
||||
0x41 0x18 0x23 0x1e
|
||||
0x41 0x18 0x63 0x1e
|
||||
|
||||
# FP16: fdiv h1, h2, h3
|
||||
# CHECK: fdiv s1, s2, s3
|
||||
# CHECK: fdiv d1, d2, d3
|
||||
|
||||
0x41 0x10 0xc3 0x1f
|
||||
0x41 0x10 0x03 0x1f
|
||||
0x41 0x10 0x43 0x1f
|
||||
|
||||
# FP16: fmadd h1, h2, h3, h4
|
||||
# CHECK: fmadd s1, s2, s3, s4
|
||||
# CHECK: fmadd d1, d2, d3, d4
|
||||
|
||||
0x41 0x48 0xe3 0x1e
|
||||
0x41 0x48 0x23 0x1e
|
||||
0x41 0x48 0x63 0x1e
|
||||
0x41 0x68 0xe3 0x1e
|
||||
0x41 0x68 0x23 0x1e
|
||||
0x41 0x68 0x63 0x1e
|
||||
|
||||
# FP16: fmax h1, h2, h3
|
||||
# CHECK: fmax s1, s2, s3
|
||||
# CHECK: fmax d1, d2, d3
|
||||
# FP16: fmaxnm h1, h2, h3
|
||||
# CHECK: fmaxnm s1, s2, s3
|
||||
# CHECK: fmaxnm d1, d2, d3
|
||||
|
||||
0x41 0x58 0xe3 0x1e
|
||||
0x41 0x58 0x23 0x1e
|
||||
0x41 0x58 0x63 0x1e
|
||||
0x41 0x78 0xe3 0x1e
|
||||
0x41 0x78 0x23 0x1e
|
||||
0x41 0x78 0x63 0x1e
|
||||
|
||||
# FP16: fmin h1, h2, h3
|
||||
# CHECK: fmin s1, s2, s3
|
||||
# CHECK: fmin d1, d2, d3
|
||||
# FP16: fminnm h1, h2, h3
|
||||
# CHECK: fminnm s1, s2, s3
|
||||
# CHECK: fminnm d1, d2, d3
|
||||
|
||||
0x41 0x90 0xc3 0x1f
|
||||
0x41 0x90 0x03 0x1f
|
||||
0x41 0x90 0x43 0x1f
|
||||
|
||||
# FP16: fmsub h1, h2, h3, h4
|
||||
# CHECK: fmsub s1, s2, s3, s4
|
||||
# CHECK: fmsub d1, d2, d3, d4
|
||||
|
||||
0x41 0x08 0xe3 0x1e
|
||||
0x41 0x08 0x23 0x1e
|
||||
0x41 0x08 0x63 0x1e
|
||||
|
||||
# FP16: fmul h1, h2, h3
|
||||
# CHECK: fmul s1, s2, s3
|
||||
# CHECK: fmul d1, d2, d3
|
||||
|
||||
0x41 0x40 0xe1 0x1e
|
||||
0x41 0x40 0x21 0x1e
|
||||
0x41 0x40 0x61 0x1e
|
||||
|
||||
# FP16: fneg h1, h2
|
||||
# CHECK: fneg s1, s2
|
||||
# CHECK: fneg d1, d2
|
||||
|
||||
0x41 0x10 0xe3 0x1f
|
||||
0x41 0x10 0x23 0x1f
|
||||
0x41 0x10 0x63 0x1f
|
||||
|
||||
# FP16: fnmadd h1, h2, h3, h4
|
||||
# CHECK: fnmadd s1, s2, s3, s4
|
||||
# CHECK: fnmadd d1, d2, d3, d4
|
||||
|
||||
0x41 0x90 0xe3 0x1f
|
||||
0x41 0x90 0x23 0x1f
|
||||
0x41 0x90 0x63 0x1f
|
||||
|
||||
# FP16: fnmsub h1, h2, h3, h4
|
||||
# CHECK: fnmsub s1, s2, s3, s4
|
||||
# CHECK: fnmsub d1, d2, d3, d4
|
||||
|
||||
0x41 0x88 0xe3 0x1e
|
||||
0x41 0x88 0x23 0x1e
|
||||
0x41 0x88 0x63 0x1e
|
||||
|
||||
# FP16: fnmul h1, h2, h3
|
||||
# CHECK: fnmul s1, s2, s3
|
||||
# CHECK: fnmul d1, d2, d3
|
||||
|
||||
0x41 0xc0 0xe1 0x1e
|
||||
0x41 0xc0 0x21 0x1e
|
||||
0x41 0xc0 0x61 0x1e
|
||||
|
||||
# FP16: fsqrt h1, h2
|
||||
# CHECK: fsqrt s1, s2
|
||||
# CHECK: fsqrt d1, d2
|
||||
|
||||
0x41 0x38 0xe3 0x1e
|
||||
0x41 0x38 0x23 0x1e
|
||||
0x41 0x38 0x63 0x1e
|
||||
|
||||
# FP16: fsub h1, h2, h3
|
||||
# CHECK: fsub s1, s2, s3
|
||||
# CHECK: fsub d1, d2, d3
|
||||
|
||||
|
|
@ -100,31 +133,43 @@
|
|||
# Floating-point comparison
|
||||
#-----------------------------------------------------------------------------
|
||||
|
||||
0x20 0x04 0xe2 0x1e
|
||||
0x20 0x04 0x22 0x1e
|
||||
0x20 0x04 0x62 0x1e
|
||||
0x30 0x04 0xe2 0x1e
|
||||
0x30 0x04 0x22 0x1e
|
||||
0x30 0x04 0x62 0x1e
|
||||
|
||||
# FP16: fccmp h1, h2, #0, eq
|
||||
# CHECK: fccmp s1, s2, #0, eq
|
||||
# CHECK: fccmp d1, d2, #0, eq
|
||||
# FP16: fccmpe h1, h2, #0, eq
|
||||
# CHECK: fccmpe s1, s2, #0, eq
|
||||
# CHECK: fccmpe d1, d2, #0, eq
|
||||
|
||||
0x20 0x20 0xe2 0x1e
|
||||
0x20 0x20 0x22 0x1e
|
||||
0x20 0x20 0x62 0x1e
|
||||
0x28 0x20 0xe0 0x1e
|
||||
0x28 0x20 0x20 0x1e
|
||||
0x28 0x20 0x60 0x1e
|
||||
0x30 0x20 0xe2 0x1e
|
||||
0x30 0x20 0x22 0x1e
|
||||
0x30 0x20 0x62 0x1e
|
||||
0x38 0x20 0xe0 0x1e
|
||||
0x38 0x20 0x20 0x1e
|
||||
0x38 0x20 0x60 0x1e
|
||||
|
||||
# FP16: fcmp h1, h2
|
||||
# CHECK: fcmp s1, s2
|
||||
# CHECK: fcmp d1, d2
|
||||
# FP16: fcmp h1, #0.0
|
||||
# CHECK: fcmp s1, #0.0
|
||||
# CHECK: fcmp d1, #0.0
|
||||
# FP16: fcmpe h1, h2
|
||||
# CHECK: fcmpe s1, s2
|
||||
# CHECK: fcmpe d1, d2
|
||||
# FP16: fcmpe h1, #0.0
|
||||
# CHECK: fcmpe s1, #0.0
|
||||
# CHECK: fcmpe d1, #0.0
|
||||
|
||||
|
|
@ -132,9 +177,11 @@
|
|||
# Floating-point conditional select
|
||||
#-----------------------------------------------------------------------------
|
||||
|
||||
0x41 0x0c 0xe3 0x1e
|
||||
0x41 0x0c 0x23 0x1e
|
||||
0x41 0x0c 0x63 0x1e
|
||||
|
||||
# FP16: fcsel h1, h2, h3, eq
|
||||
# CHECK: fcsel s1, s2, s3, eq
|
||||
# CHECK: fcsel d1, d2, d3, eq
|
||||
|
||||
|
|
@ -169,29 +216,37 @@
|
|||
# Floating-point move
|
||||
#-----------------------------------------------------------------------------
|
||||
|
||||
0x41 0x00 0xe7 0x1e
|
||||
0x41 0x00 0xe6 0x1e
|
||||
0x41 0x00 0x27 0x1e
|
||||
0x41 0x00 0x26 0x1e
|
||||
0x41 0x00 0x67 0x9e
|
||||
0x41 0x00 0x66 0x9e
|
||||
|
||||
# FP16: fmov h1, w2
|
||||
# FP16: fmov w1, h2
|
||||
# CHECK: fmov s1, w2
|
||||
# CHECK: fmov w1, s2
|
||||
# CHECK: fmov d1, x2
|
||||
# CHECK: fmov x1, d2
|
||||
|
||||
0x01 0x10 0xe8 0x1e
|
||||
0x01 0x10 0x28 0x1e
|
||||
0x01 0x10 0x68 0x1e
|
||||
0x01 0xf0 0x7b 0x1e
|
||||
0x01 0xf0 0x6b 0x1e
|
||||
|
||||
# FP16: fmov h1, #0.12500000
|
||||
# CHECK: fmov s1, #0.12500000
|
||||
# CHECK: fmov d1, #0.12500000
|
||||
# CHECK: fmov d1, #-0.48437500
|
||||
# CHECK: fmov d1, #0.48437500
|
||||
|
||||
0x41 0x40 0xe0 0x1e
|
||||
0x41 0x40 0x20 0x1e
|
||||
0x41 0x40 0x60 0x1e
|
||||
|
||||
# FP16: fmov h1, h2
|
||||
# CHECK: fmov s1, s2
|
||||
# CHECK: fmov d1, d2
|
||||
|
||||
|
|
@ -199,45 +254,59 @@
|
|||
# Floating-point round to integral
|
||||
#-----------------------------------------------------------------------------
|
||||
|
||||
0x41 0x40 0xe6 0x1e
|
||||
0x41 0x40 0x26 0x1e
|
||||
0x41 0x40 0x66 0x1e
|
||||
|
||||
# FP16: frinta h1, h2
|
||||
# CHECK: frinta s1, s2
|
||||
# CHECK: frinta d1, d2
|
||||
|
||||
0x41 0xc0 0xe7 0x1e
|
||||
0x41 0xc0 0x27 0x1e
|
||||
0x41 0xc0 0x67 0x1e
|
||||
|
||||
# FP16: frinti h1, h2
|
||||
# CHECK: frinti s1, s2
|
||||
# CHECK: frinti d1, d2
|
||||
|
||||
0x41 0x40 0xe5 0x1e
|
||||
0x41 0x40 0x25 0x1e
|
||||
0x41 0x40 0x65 0x1e
|
||||
|
||||
# FP16: frintm h1, h2
|
||||
# CHECK: frintm s1, s2
|
||||
# CHECK: frintm d1, d2
|
||||
|
||||
0x41 0x40 0xe4 0x1e
|
||||
0x41 0x40 0x24 0x1e
|
||||
0x41 0x40 0x64 0x1e
|
||||
|
||||
# FP16: frintn h1, h2
|
||||
# CHECK: frintn s1, s2
|
||||
# CHECK: frintn d1, d2
|
||||
|
||||
0x41 0xc0 0xe4 0x1e
|
||||
0x41 0xc0 0x24 0x1e
|
||||
0x41 0xc0 0x64 0x1e
|
||||
|
||||
# FP16: frintp h1, h2
|
||||
# CHECK: frintp s1, s2
|
||||
# CHECK: frintp d1, d2
|
||||
|
||||
0x41 0x40 0xe7 0x1e
|
||||
0x41 0x40 0x27 0x1e
|
||||
0x41 0x40 0x67 0x1e
|
||||
|
||||
# FP16: frintx h1, h2
|
||||
# CHECK: frintx s1, s2
|
||||
# CHECK: frintx d1, d2
|
||||
|
||||
0x41 0xc0 0xe5 0x1e
|
||||
0x41 0xc0 0x25 0x1e
|
||||
0x41 0xc0 0x65 0x1e
|
||||
|
||||
# FP16: frintz h1, h2
|
||||
# CHECK: frintz s1, s2
|
||||
# CHECK: frintz d1, d2
|
||||
|
||||
|
|
|
|||
|
|
@ -1,5 +1,6 @@
|
|||
# RUN: llvm-mc -triple=aarch64 -mattr=+fp-armv8 -disassemble < %s | FileCheck %s
|
||||
# RUN: llvm-mc -triple=arm64 -mattr=+fp-armv8 -disassemble < %s | FileCheck %s
|
||||
# RUN: llvm-mc -triple=arm64 -mattr=+fp-armv8,+fullfp16 -disassemble < %s | FileCheck %s --check-prefix=CHECK --check-prefix=FP16
|
||||
|
||||
#------------------------------------------------------------------------------
|
||||
# Add/sub (immediate)
|
||||
|
|
@ -1516,6 +1517,20 @@
|
|||
# Floating-point <-> fixed-point conversion
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
# FP16: fcvtzs w3, h5, #1
|
||||
# FP16: fcvtzs wzr, h20, #13
|
||||
# FP16: fcvtzs w19, h0, #32
|
||||
0xa3 0xfc 0xd8 0x1e
|
||||
0x9f 0xce 0xd8 0x1e
|
||||
0x13 0x80 0xd8 0x1e
|
||||
|
||||
# FP16: fcvtzs x3, h5, #1
|
||||
# FP16: fcvtzs x12, h30, #45
|
||||
# FP16: fcvtzs x19, h0, #64
|
||||
0xa3 0xfc 0xd8 0x9e
|
||||
0xcc 0x4f 0xd8 0x9e
|
||||
0x13 0x00 0xd8 0x9e
|
||||
|
||||
# CHECK: fcvtzs w3, s5, #1
|
||||
# CHECK: fcvtzs wzr, s20, #13
|
||||
# CHECK: fcvtzs w19, s0, #32
|
||||
|
|
@ -1544,6 +1559,20 @@
|
|||
0xcc 0x4f 0x58 0x9e
|
||||
0x13 0x00 0x58 0x9e
|
||||
|
||||
# FP16: fcvtzu w3, h5, #1
|
||||
# FP16: fcvtzu wzr, h20, #13
|
||||
# FP16: fcvtzu w19, h0, #32
|
||||
0xa3 0xfc 0xd9 0x1e
|
||||
0x9f 0xce 0xd9 0x1e
|
||||
0x13 0x80 0xd9 0x1e
|
||||
|
||||
# FP16: fcvtzu x3, h5, #1
|
||||
# FP16: fcvtzu x12, h30, #45
|
||||
# FP16: fcvtzu x19, h0, #64
|
||||
0xa3 0xfc 0xd9 0x9e
|
||||
0xcc 0x4f 0xd9 0x9e
|
||||
0x13 0x00 0xd9 0x9e
|
||||
|
||||
# CHECK: fcvtzu w3, s5, #1
|
||||
# CHECK: fcvtzu wzr, s20, #13
|
||||
# CHECK: fcvtzu w19, s0, #32
|
||||
|
|
@ -1572,6 +1601,20 @@
|
|||
0xcc 0x4f 0x59 0x9e
|
||||
0x13 0x00 0x59 0x9e
|
||||
|
||||
# FP16: scvtf h23, w19, #1
|
||||
# FP16: scvtf h31, wzr, #20
|
||||
# FP16: scvtf h14, w0, #32
|
||||
0x77 0xfe 0xc2 0x1e
|
||||
0xff 0xb3 0xc2 0x1e
|
||||
0x0e 0x80 0xc2 0x1e
|
||||
|
||||
# FP16: scvtf h23, x19, #1
|
||||
# FP16: scvtf h31, xzr, #20
|
||||
# FP16: scvtf h14, x0, #64
|
||||
0x77 0xfe 0xc2 0x9e
|
||||
0xff 0xb3 0xc2 0x9e
|
||||
0x0e 0x00 0xc2 0x9e
|
||||
|
||||
# CHECK: scvtf s23, w19, #1
|
||||
# CHECK: scvtf s31, wzr, #20
|
||||
# CHECK: scvtf s14, w0, #32
|
||||
|
|
@ -1600,6 +1643,20 @@
|
|||
0xff 0xb3 0x42 0x9e
|
||||
0x0e 0x00 0x42 0x9e
|
||||
|
||||
# FP16: ucvtf h23, w19, #1
|
||||
# FP16: ucvtf h31, wzr, #20
|
||||
# FP16: ucvtf h14, w0, #32
|
||||
0x77 0xfe 0xc3 0x1e
|
||||
0xff 0xb3 0xc3 0x1e
|
||||
0x0e 0x80 0xc3 0x1e
|
||||
|
||||
# FP16: ucvtf h23, x19, #1
|
||||
# FP16: ucvtf h31, xzr, #20
|
||||
# FP16: ucvtf h14, x0, #64
|
||||
0x77 0xfe 0xc3 0x9e
|
||||
0xff 0xb3 0xc3 0x9e
|
||||
0x0e 0x00 0xc3 0x9e
|
||||
|
||||
# CHECK: ucvtf s23, w19, #1
|
||||
# CHECK: ucvtf s31, wzr, #20
|
||||
# CHECK: ucvtf s14, w0, #32
|
||||
|
|
@ -1631,6 +1688,61 @@
|
|||
#------------------------------------------------------------------------------
|
||||
# Floating-point <-> integer conversion
|
||||
#------------------------------------------------------------------------------
|
||||
|
||||
# FP16: fcvtns w3, h31
|
||||
# FP16: fcvtns xzr, h12
|
||||
# FP16: fcvtnu wzr, h12
|
||||
# FP16: fcvtnu x0, h0
|
||||
0xe3 0x3 0xe0 0x1e
|
||||
0x9f 0x1 0xe0 0x9e
|
||||
0x9f 0x1 0xe1 0x1e
|
||||
0x0 0x0 0xe1 0x9e
|
||||
|
||||
# FP16: fcvtps wzr, h9
|
||||
# FP16: fcvtps x12, h20
|
||||
# FP16: fcvtpu w30, h23
|
||||
# FP16: fcvtpu x29, h3
|
||||
0x3f 0x1 0xe8 0x1e
|
||||
0x8c 0x2 0xe8 0x9e
|
||||
0xfe 0x2 0xe9 0x1e
|
||||
0x7d 0x0 0xe9 0x9e
|
||||
|
||||
# FP16: fcvtms w2, h3
|
||||
# FP16: fcvtms x4, h5
|
||||
# FP16: fcvtmu w6, h7
|
||||
# FP16: fcvtmu x8, h9
|
||||
0x62 0x0 0xf0 0x1e
|
||||
0xa4 0x0 0xf0 0x9e
|
||||
0xe6 0x0 0xf1 0x1e
|
||||
0x28 0x1 0xf1 0x9e
|
||||
|
||||
# FP16: fcvtzs w10, h11
|
||||
# FP16: fcvtzs x12, h13
|
||||
# FP16: fcvtzu w14, h15
|
||||
# FP16: fcvtzu x15, h16
|
||||
0x6a 0x1 0xf8 0x1e
|
||||
0xac 0x1 0xf8 0x9e
|
||||
0xee 0x1 0xf9 0x1e
|
||||
0xf 0x2 0xf9 0x9e
|
||||
|
||||
# FP16: scvtf h17, w18
|
||||
# FP16: scvtf h19, x20
|
||||
# FP16: ucvtf h21, w22
|
||||
# FP16: scvtf h23, x24
|
||||
0x51 0x2 0xe2 0x1e
|
||||
0x93 0x2 0xe2 0x9e
|
||||
0xd5 0x2 0xe3 0x1e
|
||||
0x17 0x3 0xe2 0x9e
|
||||
|
||||
# FP16: fcvtas w25, h26
|
||||
# FP16: fcvtas x27, h28
|
||||
# FP16: fcvtau w29, h30
|
||||
# FP16: fcvtau xzr, h0
|
||||
0x59 0x3 0xe4 0x1e
|
||||
0x9b 0x3 0xe4 0x9e
|
||||
0xdd 0x3 0xe5 0x1e
|
||||
0x1f 0x0 0xe5 0x9e
|
||||
|
||||
# CHECK: fcvtns w3, s31
|
||||
# CHECK: fcvtns xzr, s12
|
||||
# CHECK: fcvtnu wzr, s12
|
||||
|
|
|
|||
|
|
@ -0,0 +1,145 @@
|
|||
# RUN: not llvm-mc -disassemble -triple=aarch64 -mattr=+neon,-fullfp16 < %s 2>&1 | FileCheck %s
|
||||
# RUN: not llvm-mc -disassemble -triple=aarch64 -mattr=-neon,-fullfp16 < %s 2>&1 | FileCheck %s
|
||||
|
||||
[0x41,0xc0,0xe0,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x28,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x18,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x10,0xc3,0x1f]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x48,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x68,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x58,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x78,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x90,0xc3,0x1f]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x08,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x40,0xe1,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x10,0xe3,0x1f]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x90,0xe3,0x1f]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x88,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xc0,0xe1,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x38,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x20,0x04,0xe2,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x30,0x04,0xe2,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x20,0x20,0xe2,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x28,0x20,0xe0,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x30,0x20,0xe2,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x38,0x20,0xe0,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x0c,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe4,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe4,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe5,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe5,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf0,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf0,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf1,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf1,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe0,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe0,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe1,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe1,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe8,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe8,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe9,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe9,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf8,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xd8,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf8,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xd8,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf9,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xd9,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xf9,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xd9,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe2,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xc2,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe2,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xc2,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xc3,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe3,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xfc,0xc3,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe7,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe6,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe7,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x00,0xe6,0x9e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x01,0x10,0xe8,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x01,0x10,0xe8,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0xe2,0x03,0xe7,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x40,0xe0,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x40,0xe6,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xc0,0xe7,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x40,0xe5,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x40,0xe4,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xc0,0xe4,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0x40,0xe7,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
[0x41,0xc0,0xe5,0x1e]
|
||||
# CHECK: warning: invalid instruction encoding
|
||||
|
||||
# CHECK-NOT: warning: invalid instruction encoding
|
||||
Loading…
Reference in New Issue