221270db90
* More accurate impl of FMINNM/FMAXNM, add vector variants * Optimize for the 0 case when op1 != op2 * Address PR feedback
561 lines
19 KiB
C#
561 lines
19 KiB
C#
using ChocolArm64.State;
|
|
using ChocolArm64.Translation;
|
|
using System;
|
|
using System.Runtime.CompilerServices;
|
|
using System.Runtime.Intrinsics;
|
|
using System.Runtime.Intrinsics.X86;
|
|
|
|
namespace ChocolArm64.Instruction
|
|
{
|
|
static class AVectorHelper
|
|
{
|
|
public static void EmitCall(AILEmitterCtx Context, string Name64, string Name128)
|
|
{
|
|
bool IsSimd64 = Context.CurrOp.RegisterSize == ARegisterSize.SIMD64;
|
|
|
|
Context.EmitCall(typeof(AVectorHelper), IsSimd64 ? Name64 : Name128);
|
|
}
|
|
|
|
public static void EmitCall(AILEmitterCtx Context, string MthdName)
|
|
{
|
|
Context.EmitCall(typeof(AVectorHelper), MthdName);
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static int SatF32ToS32(float Value)
|
|
{
|
|
if (float.IsNaN(Value)) return 0;
|
|
|
|
return Value > int.MaxValue ? int.MaxValue :
|
|
Value < int.MinValue ? int.MinValue : (int)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static long SatF32ToS64(float Value)
|
|
{
|
|
if (float.IsNaN(Value)) return 0;
|
|
|
|
return Value > long.MaxValue ? long.MaxValue :
|
|
Value < long.MinValue ? long.MinValue : (long)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static uint SatF32ToU32(float Value)
|
|
{
|
|
if (float.IsNaN(Value)) return 0;
|
|
|
|
return Value > uint.MaxValue ? uint.MaxValue :
|
|
Value < uint.MinValue ? uint.MinValue : (uint)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static ulong SatF32ToU64(float Value)
|
|
{
|
|
if (float.IsNaN(Value)) return 0;
|
|
|
|
return Value > ulong.MaxValue ? ulong.MaxValue :
|
|
Value < ulong.MinValue ? ulong.MinValue : (ulong)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static int SatF64ToS32(double Value)
|
|
{
|
|
if (double.IsNaN(Value)) return 0;
|
|
|
|
return Value > int.MaxValue ? int.MaxValue :
|
|
Value < int.MinValue ? int.MinValue : (int)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static long SatF64ToS64(double Value)
|
|
{
|
|
if (double.IsNaN(Value)) return 0;
|
|
|
|
return Value > long.MaxValue ? long.MaxValue :
|
|
Value < long.MinValue ? long.MinValue : (long)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static uint SatF64ToU32(double Value)
|
|
{
|
|
if (double.IsNaN(Value)) return 0;
|
|
|
|
return Value > uint.MaxValue ? uint.MaxValue :
|
|
Value < uint.MinValue ? uint.MinValue : (uint)Value;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static ulong SatF64ToU64(double Value)
|
|
{
|
|
if (double.IsNaN(Value)) return 0;
|
|
|
|
return Value > ulong.MaxValue ? ulong.MaxValue :
|
|
Value < ulong.MinValue ? ulong.MinValue : (ulong)Value;
|
|
}
|
|
|
|
public static double Round(double Value, int Fpcr)
|
|
{
|
|
switch ((ARoundMode)((Fpcr >> 22) & 3))
|
|
{
|
|
case ARoundMode.ToNearest: return Math.Round (Value);
|
|
case ARoundMode.TowardsPlusInfinity: return Math.Ceiling (Value);
|
|
case ARoundMode.TowardsMinusInfinity: return Math.Floor (Value);
|
|
case ARoundMode.TowardsZero: return Math.Truncate(Value);
|
|
}
|
|
|
|
throw new InvalidOperationException();
|
|
}
|
|
|
|
public static float RoundF(float Value, int Fpcr)
|
|
{
|
|
switch ((ARoundMode)((Fpcr >> 22) & 3))
|
|
{
|
|
case ARoundMode.ToNearest: return MathF.Round (Value);
|
|
case ARoundMode.TowardsPlusInfinity: return MathF.Ceiling (Value);
|
|
case ARoundMode.TowardsMinusInfinity: return MathF.Floor (Value);
|
|
case ARoundMode.TowardsZero: return MathF.Truncate(Value);
|
|
}
|
|
|
|
throw new InvalidOperationException();
|
|
}
|
|
|
|
public static Vector128<float> Tbl1_V64(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0)
|
|
{
|
|
return Tbl(Vector, 8, Tb0);
|
|
}
|
|
|
|
public static Vector128<float> Tbl1_V128(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0)
|
|
{
|
|
return Tbl(Vector, 16, Tb0);
|
|
}
|
|
|
|
public static Vector128<float> Tbl2_V64(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0,
|
|
Vector128<float> Tb1)
|
|
{
|
|
return Tbl(Vector, 8, Tb0, Tb1);
|
|
}
|
|
|
|
public static Vector128<float> Tbl2_V128(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0,
|
|
Vector128<float> Tb1)
|
|
{
|
|
return Tbl(Vector, 16, Tb0, Tb1);
|
|
}
|
|
|
|
public static Vector128<float> Tbl3_V64(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0,
|
|
Vector128<float> Tb1,
|
|
Vector128<float> Tb2)
|
|
{
|
|
return Tbl(Vector, 8, Tb0, Tb1, Tb2);
|
|
}
|
|
|
|
public static Vector128<float> Tbl3_V128(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0,
|
|
Vector128<float> Tb1,
|
|
Vector128<float> Tb2)
|
|
{
|
|
return Tbl(Vector, 16, Tb0, Tb1, Tb2);
|
|
}
|
|
|
|
public static Vector128<float> Tbl4_V64(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0,
|
|
Vector128<float> Tb1,
|
|
Vector128<float> Tb2,
|
|
Vector128<float> Tb3)
|
|
{
|
|
return Tbl(Vector, 8, Tb0, Tb1, Tb2, Tb3);
|
|
}
|
|
|
|
public static Vector128<float> Tbl4_V128(
|
|
Vector128<float> Vector,
|
|
Vector128<float> Tb0,
|
|
Vector128<float> Tb1,
|
|
Vector128<float> Tb2,
|
|
Vector128<float> Tb3)
|
|
{
|
|
return Tbl(Vector, 16, Tb0, Tb1, Tb2, Tb3);
|
|
}
|
|
|
|
private static Vector128<float> Tbl(Vector128<float> Vector, int Bytes, params Vector128<float>[] Tb)
|
|
{
|
|
Vector128<float> Res = new Vector128<float>();
|
|
|
|
byte[] Table = new byte[Tb.Length * 16];
|
|
|
|
for (byte Index = 0; Index < Tb.Length; Index++)
|
|
for (byte Index2 = 0; Index2 < 16; Index2++)
|
|
{
|
|
Table[Index * 16 + Index2] = (byte)VectorExtractIntZx(Tb[Index], Index2, 0);
|
|
}
|
|
|
|
for (byte Index = 0; Index < Bytes; Index++)
|
|
{
|
|
byte TblIdx = (byte)VectorExtractIntZx(Vector, Index, 0);
|
|
|
|
if (TblIdx < Table.Length)
|
|
{
|
|
Res = VectorInsertInt(Table[TblIdx], Res, Index, 0);
|
|
}
|
|
}
|
|
|
|
return Res;
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static double VectorExtractDouble(Vector128<float> Vector, byte Index)
|
|
{
|
|
return BitConverter.Int64BitsToDouble(VectorExtractIntSx(Vector, Index, 3));
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static long VectorExtractIntSx(Vector128<float> Vector, byte Index, int Size)
|
|
{
|
|
if (Sse41.IsSupported)
|
|
{
|
|
switch (Size)
|
|
{
|
|
case 0:
|
|
return (sbyte)Sse41.Extract(Sse.StaticCast<float, byte>(Vector), Index);
|
|
|
|
case 1:
|
|
return (short)Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), Index);
|
|
|
|
case 2:
|
|
return Sse41.Extract(Sse.StaticCast<float, int>(Vector), Index);
|
|
|
|
case 3:
|
|
return Sse41.Extract(Sse.StaticCast<float, long>(Vector), Index);
|
|
}
|
|
|
|
throw new ArgumentOutOfRangeException(nameof(Size));
|
|
}
|
|
else if (Sse2.IsSupported)
|
|
{
|
|
switch (Size)
|
|
{
|
|
case 0:
|
|
return (sbyte)VectorExtractIntZx(Vector, Index, Size);
|
|
|
|
case 1:
|
|
return (short)VectorExtractIntZx(Vector, Index, Size);
|
|
|
|
case 2:
|
|
return (int)VectorExtractIntZx(Vector, Index, Size);
|
|
|
|
case 3:
|
|
return (long)VectorExtractIntZx(Vector, Index, Size);
|
|
}
|
|
|
|
throw new ArgumentOutOfRangeException(nameof(Size));
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static ulong VectorExtractIntZx(Vector128<float> Vector, byte Index, int Size)
|
|
{
|
|
if (Sse41.IsSupported)
|
|
{
|
|
switch (Size)
|
|
{
|
|
case 0:
|
|
return Sse41.Extract(Sse.StaticCast<float, byte>(Vector), Index);
|
|
|
|
case 1:
|
|
return Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), Index);
|
|
|
|
case 2:
|
|
return Sse41.Extract(Sse.StaticCast<float, uint>(Vector), Index);
|
|
|
|
case 3:
|
|
return Sse41.Extract(Sse.StaticCast<float, ulong>(Vector), Index);
|
|
}
|
|
|
|
throw new ArgumentOutOfRangeException(nameof(Size));
|
|
}
|
|
else if (Sse2.IsSupported)
|
|
{
|
|
int ShortIdx = Size == 0
|
|
? Index >> 1
|
|
: Index << (Size - 1);
|
|
|
|
ushort Value = Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), (byte)ShortIdx);
|
|
|
|
switch (Size)
|
|
{
|
|
case 0:
|
|
return (byte)(Value >> (Index & 1) * 8);
|
|
|
|
case 1:
|
|
return Value;
|
|
|
|
case 2:
|
|
case 3:
|
|
{
|
|
ushort Value1 = Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), (byte)(ShortIdx + 1));
|
|
|
|
if (Size == 2)
|
|
{
|
|
return (uint)(Value | (Value1 << 16));
|
|
}
|
|
|
|
ushort Value2 = Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), (byte)(ShortIdx + 2));
|
|
ushort Value3 = Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), (byte)(ShortIdx + 3));
|
|
|
|
return ((ulong)Value << 0) |
|
|
((ulong)Value1 << 16) |
|
|
((ulong)Value2 << 32) |
|
|
((ulong)Value3 << 48);
|
|
}
|
|
}
|
|
|
|
throw new ArgumentOutOfRangeException(nameof(Size));
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static float VectorExtractSingle(Vector128<float> Vector, byte Index)
|
|
{
|
|
if (Sse41.IsSupported)
|
|
{
|
|
return Sse41.Extract(Vector, Index);
|
|
}
|
|
else if (Sse2.IsSupported)
|
|
{
|
|
Vector128<ushort> ShortVector = Sse.StaticCast<float, ushort>(Vector);
|
|
|
|
int Low = Sse2.Extract(ShortVector, (byte)(Index * 2 + 0));
|
|
int High = Sse2.Extract(ShortVector, (byte)(Index * 2 + 1));
|
|
|
|
return BitConverter.Int32BitsToSingle(Low | (High << 16));
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorInsertDouble(double Value, Vector128<float> Vector, byte Index)
|
|
{
|
|
return VectorInsertInt((ulong)BitConverter.DoubleToInt64Bits(Value), Vector, Index, 3);
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorInsertInt(ulong Value, Vector128<float> Vector, byte Index, int Size)
|
|
{
|
|
if (Sse41.IsSupported)
|
|
{
|
|
switch (Size)
|
|
{
|
|
case 0:
|
|
return Sse.StaticCast<byte, float>(Sse41.Insert(Sse.StaticCast<float, byte>(Vector), (byte)Value, Index));
|
|
|
|
case 1:
|
|
return Sse.StaticCast<ushort, float>(Sse2.Insert(Sse.StaticCast<float, ushort>(Vector), (ushort)Value, Index));
|
|
|
|
case 2:
|
|
return Sse.StaticCast<uint, float>(Sse41.Insert(Sse.StaticCast<float, uint>(Vector), (uint)Value, Index));
|
|
|
|
case 3:
|
|
return Sse.StaticCast<ulong, float>(Sse41.Insert(Sse.StaticCast<float, ulong>(Vector), Value, Index));
|
|
}
|
|
|
|
throw new ArgumentOutOfRangeException(nameof(Size));
|
|
}
|
|
else if (Sse2.IsSupported)
|
|
{
|
|
Vector128<ushort> ShortVector = Sse.StaticCast<float, ushort>(Vector);
|
|
|
|
int ShortIdx = Size == 0
|
|
? Index >> 1
|
|
: Index << (Size - 1);
|
|
|
|
switch (Size)
|
|
{
|
|
case 0:
|
|
{
|
|
ushort ShortVal = Sse2.Extract(Sse.StaticCast<float, ushort>(Vector), (byte)ShortIdx);
|
|
|
|
int Shift = (Index & 1) * 8;
|
|
|
|
ShortVal &= (ushort)(0xff00 >> Shift);
|
|
|
|
ShortVal |= (ushort)((byte)Value << Shift);
|
|
|
|
return Sse.StaticCast<ushort, float>(Sse2.Insert(ShortVector, ShortVal, (byte)ShortIdx));
|
|
}
|
|
|
|
case 1:
|
|
return Sse.StaticCast<ushort, float>(Sse2.Insert(Sse.StaticCast<float, ushort>(Vector), (ushort)Value, Index));
|
|
|
|
case 2:
|
|
case 3:
|
|
{
|
|
ShortVector = Sse2.Insert(ShortVector, (ushort)(Value >> 0), (byte)(ShortIdx + 0));
|
|
ShortVector = Sse2.Insert(ShortVector, (ushort)(Value >> 16), (byte)(ShortIdx + 1));
|
|
|
|
if (Size == 3)
|
|
{
|
|
ShortVector = Sse2.Insert(ShortVector, (ushort)(Value >> 32), (byte)(ShortIdx + 2));
|
|
ShortVector = Sse2.Insert(ShortVector, (ushort)(Value >> 48), (byte)(ShortIdx + 3));
|
|
}
|
|
|
|
return Sse.StaticCast<ushort, float>(ShortVector);
|
|
}
|
|
}
|
|
|
|
throw new ArgumentOutOfRangeException(nameof(Size));
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorInsertSingle(float Value, Vector128<float> Vector, byte Index)
|
|
{
|
|
if (Sse41.IsSupported)
|
|
{
|
|
return Sse41.Insert(Vector, Value, (byte)(Index << 4));
|
|
}
|
|
else if (Sse2.IsSupported)
|
|
{
|
|
int IntValue = BitConverter.SingleToInt32Bits(Value);
|
|
|
|
ushort Low = (ushort)(IntValue >> 0);
|
|
ushort High = (ushort)(IntValue >> 16);
|
|
|
|
Vector128<ushort> ShortVector = Sse.StaticCast<float, ushort>(Vector);
|
|
|
|
ShortVector = Sse2.Insert(ShortVector, Low, (byte)(Index * 2 + 0));
|
|
ShortVector = Sse2.Insert(ShortVector, High, (byte)(Index * 2 + 1));
|
|
|
|
return Sse.StaticCast<ushort, float>(ShortVector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<sbyte> VectorSingleToSByte(Vector128<float> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<float, sbyte>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<short> VectorSingleToInt16(Vector128<float> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<float, short>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<int> VectorSingleToInt32(Vector128<float> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<float, int>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<long> VectorSingleToInt64(Vector128<float> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<float, long>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<double> VectorSingleToDouble(Vector128<float> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<float, double>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorSByteToSingle(Vector128<sbyte> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<sbyte, float>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorInt16ToSingle(Vector128<short> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<short, float>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorInt32ToSingle(Vector128<int> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<int, float>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorInt64ToSingle(Vector128<long> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<long, float>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
public static Vector128<float> VectorDoubleToSingle(Vector128<double> Vector)
|
|
{
|
|
if (Sse.IsSupported)
|
|
{
|
|
return Sse.StaticCast<double, float>(Vector);
|
|
}
|
|
|
|
throw new PlatformNotSupportedException();
|
|
}
|
|
}
|
|
}
|