Fix Vnmls_S fast path (F64: losing input d value). Fix Vnmla_S & Vnmls_S slow paths (using fused inst.s). Fix Vfma_V slow path not using StandardFPSCRValue(). (#1775)
* Fix Vnmls_S fast path (F64: losing input d value). Fix Vnmla_S & Vnmls_S slow paths (using fused inst.s). Add Vfma_S & Vfms_S Fma fast paths. Add Vfnma_S inst. with Fma/Sse fast paths and slow path. Add Vfnms_S Sse fast path. Add Tests for affected inst.s. Nits. * InternalVersion = 1775 * Nits. * Fix Vfma_V slow path not using StandardFPSCRValue(). * Nit: Fix Vfma_V order. * Add Vfms_V Sse fast path and slow path. * Add Vfma_V and Vfms_V Test.
This commit is contained in:
parent
b5c215111d
commit
8a33e884f8
13 changed files with 292 additions and 221 deletions
|
@ -274,17 +274,15 @@ namespace ARMeilleure.CodeGen.X86
|
|||
Add(X86Instruction.Vcvtph2ps, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f3813, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vcvtps2ph, new InstructionInfo(0x000f3a1d, BadOp, BadOp, BadOp, BadOp, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmadd231ps, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38b8, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmadd231pd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38b8, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfmadd231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38b9, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmadd231sd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38b9, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfmsub231ps, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38ba, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmsub231pd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38ba, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfmsub231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bb, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmadd231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38b9, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmsub231sd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bb, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfnmsub231ps, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38be, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfnmsub231pd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38be, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfnmsub231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bf, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfmsub231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bb, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfnmadd231ps, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bc, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfnmadd231sd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bd, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfnmadd231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bd, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vfnmsub231sd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bf, InstructionFlags.Vex | InstructionFlags.Prefix66 | InstructionFlags.RexW));
|
||||
Add(X86Instruction.Vfnmsub231ss, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f38bf, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Vpblendvb, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x000f3a4c, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
Add(X86Instruction.Xor, new InstructionInfo(0x00000031, 0x06000083, 0x06000081, BadOp, 0x00000033, InstructionFlags.None));
|
||||
Add(X86Instruction.Xorpd, new InstructionInfo(BadOp, BadOp, BadOp, BadOp, 0x00000f57, InstructionFlags.Vex | InstructionFlags.Prefix66));
|
||||
|
|
|
@ -440,9 +440,12 @@ namespace ARMeilleure.CodeGen.X86
|
|||
else
|
||||
{
|
||||
EnsureSameReg(dest, src1);
|
||||
|
||||
Debug.Assert(src3.GetRegister().Index == 0);
|
||||
|
||||
context.Assembler.WriteInstruction(info.Inst, dest, src1, src2);
|
||||
}
|
||||
|
||||
break;
|
||||
}
|
||||
|
||||
|
@ -474,11 +477,16 @@ namespace ARMeilleure.CodeGen.X86
|
|||
Operand src2 = operation.GetSource(1);
|
||||
Operand src3 = operation.GetSource(2);
|
||||
|
||||
EnsureSameType(dest, src1, src2, src3);
|
||||
EnsureSameReg(dest, src1);
|
||||
Debug.Assert(!dest.Type.IsInteger());
|
||||
Debug.Assert(HardwareCapabilities.SupportsVexEncoding);
|
||||
|
||||
Debug.Assert(dest.Kind == OperandKind.Register && src1.Kind == OperandKind.Register && src2.Kind == OperandKind.Register);
|
||||
Debug.Assert(src3.Kind == OperandKind.Register || src3.Kind == OperandKind.Memory);
|
||||
|
||||
EnsureSameType(dest, src1, src2, src3);
|
||||
Debug.Assert(dest.Type == OperandType.V128);
|
||||
|
||||
Debug.Assert(dest.Value == src1.Value);
|
||||
|
||||
context.Assembler.WriteInstruction(info.Inst, dest, src2, src3);
|
||||
|
||||
break;
|
||||
|
|
|
@ -166,16 +166,14 @@ namespace ARMeilleure.CodeGen.X86
|
|||
Add(Intrinsic.X86Unpcklps, new IntrinsicInfo(X86Instruction.Unpcklps, IntrinsicType.Binary));
|
||||
Add(Intrinsic.X86Vcvtph2ps, new IntrinsicInfo(X86Instruction.Vcvtph2ps, IntrinsicType.Unary));
|
||||
Add(Intrinsic.X86Vcvtps2ph, new IntrinsicInfo(X86Instruction.Vcvtps2ph, IntrinsicType.BinaryImm));
|
||||
Add(Intrinsic.X86Vfmadd231pd, new IntrinsicInfo(X86Instruction.Vfmadd231pd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmadd231ps, new IntrinsicInfo(X86Instruction.Vfmadd231ps, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmadd231sd, new IntrinsicInfo(X86Instruction.Vfmadd231sd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmadd231ss, new IntrinsicInfo(X86Instruction.Vfmadd231ss, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmsub231pd, new IntrinsicInfo(X86Instruction.Vfmsub231pd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmsub231ps, new IntrinsicInfo(X86Instruction.Vfmsub231ps, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmsub231sd, new IntrinsicInfo(X86Instruction.Vfmsub231sd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfmsub231ss, new IntrinsicInfo(X86Instruction.Vfmsub231ss, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmsub231pd, new IntrinsicInfo(X86Instruction.Vfnmsub231pd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmsub231ps, new IntrinsicInfo(X86Instruction.Vfnmsub231ps, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmadd231ps, new IntrinsicInfo(X86Instruction.Vfnmadd231ps, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmadd231sd, new IntrinsicInfo(X86Instruction.Vfnmadd231sd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmadd231ss, new IntrinsicInfo(X86Instruction.Vfnmadd231ss, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmsub231sd, new IntrinsicInfo(X86Instruction.Vfnmsub231sd, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Vfnmsub231ss, new IntrinsicInfo(X86Instruction.Vfnmsub231ss, IntrinsicType.Fma));
|
||||
Add(Intrinsic.X86Xorpd, new IntrinsicInfo(X86Instruction.Xorpd, IntrinsicType.Binary));
|
||||
|
|
|
@ -203,18 +203,16 @@ namespace ARMeilleure.CodeGen.X86
|
|||
Vblendvps,
|
||||
Vcvtph2ps,
|
||||
Vcvtps2ph,
|
||||
Vfmadd231pd,
|
||||
Vfmadd231ps,
|
||||
Vfmadd231sd,
|
||||
Vfmadd231ss,
|
||||
Vfmsub231ps,
|
||||
Vfmsub231pd,
|
||||
Vfmsub231ss,
|
||||
Vfmsub231sd,
|
||||
Vfnmsub231ps,
|
||||
Vfnmsub231pd,
|
||||
Vfnmsub231ss,
|
||||
Vfmsub231ss,
|
||||
Vfnmadd231ps,
|
||||
Vfnmadd231sd,
|
||||
Vfnmadd231ss,
|
||||
Vfnmsub231sd,
|
||||
Vfnmsub231ss,
|
||||
Vpblendvb,
|
||||
Xor,
|
||||
Xorpd,
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue