mirror of
https://github.com/Ryujinx/Ryujinx.git
synced 2024-08-18 14:48:44 +02:00
* infra: Migrate to .NET 5 This migrate projects and CI to .NET 5 * Remove language version restrictions (now on 9.0 by default) * infra: pin .NET 5 to avoid later issues * infra: Cleanup csproj files * infra: update dependencies * infra: Add temporary workaround for a bug in Vector128.Create see https://github.com/dotnet/runtime/issues/44704 for more informations
187 lines
7.6 KiB
C#
187 lines
7.6 KiB
C#
using Ryujinx.Graphics.Vic.Image;
|
|
using Ryujinx.Graphics.Vic.Types;
|
|
using System;
|
|
using System.Diagnostics;
|
|
using System.Runtime.CompilerServices;
|
|
using System.Runtime.Intrinsics;
|
|
using System.Runtime.Intrinsics.X86;
|
|
|
|
namespace Ryujinx.Graphics.Vic
|
|
{
|
|
static class Blender
|
|
{
|
|
public static void BlendOne(Surface dst, Surface src, ref SlotStruct slot)
|
|
{
|
|
if (Sse41.IsSupported && (dst.Width & 3) == 0)
|
|
{
|
|
BlendOneSse41(dst, src, ref slot);
|
|
return;
|
|
}
|
|
|
|
for (int y = 0; y < dst.Height; y++)
|
|
{
|
|
for (int x = 0; x < dst.Width; x++)
|
|
{
|
|
int inR = src.GetR(x, y);
|
|
int inG = src.GetG(x, y);
|
|
int inB = src.GetB(x, y);
|
|
|
|
MatrixMultiply(ref slot.ColorMatrixStruct, inR, inG, inB, out int r, out int g, out int b);
|
|
|
|
r = Math.Clamp(r, slot.SlotConfig.SoftClampLow, slot.SlotConfig.SoftClampHigh);
|
|
g = Math.Clamp(g, slot.SlotConfig.SoftClampLow, slot.SlotConfig.SoftClampHigh);
|
|
b = Math.Clamp(b, slot.SlotConfig.SoftClampLow, slot.SlotConfig.SoftClampHigh);
|
|
|
|
dst.SetR(x, y, (ushort)r);
|
|
dst.SetG(x, y, (ushort)g);
|
|
dst.SetB(x, y, (ushort)b);
|
|
dst.SetA(x, y, src.GetA(x, y));
|
|
}
|
|
}
|
|
}
|
|
|
|
private unsafe static void BlendOneSse41(Surface dst, Surface src, ref SlotStruct slot)
|
|
{
|
|
Debug.Assert((dst.Width & 3) == 0);
|
|
|
|
ref MatrixStruct mtx = ref slot.ColorMatrixStruct;
|
|
|
|
int one = 1 << (mtx.MatrixRShift + 8);
|
|
|
|
|
|
// NOTE: This is buggy on .NET 5.0.100, we use a workaround for now (see https://github.com/dotnet/runtime/issues/44704)
|
|
// TODO: Uncomment this when fixed.
|
|
//Vector128<int> col1 = Vector128.Create(mtx.MatrixCoeff00, mtx.MatrixCoeff10, mtx.MatrixCoeff20, 0);
|
|
//Vector128<int> col2 = Vector128.Create(mtx.MatrixCoeff01, mtx.MatrixCoeff11, mtx.MatrixCoeff21, 0);
|
|
//Vector128<int> col3 = Vector128.Create(mtx.MatrixCoeff02, mtx.MatrixCoeff12, mtx.MatrixCoeff22, one);
|
|
//Vector128<int> col4 = Vector128.Create(mtx.MatrixCoeff03, mtx.MatrixCoeff13, mtx.MatrixCoeff23, 0);
|
|
|
|
Vector128<int> col1 = new Vector128<int>();
|
|
Vector128<int> col2 = new Vector128<int>();
|
|
Vector128<int> col3 = new Vector128<int>();
|
|
Vector128<int> col4 = new Vector128<int>();
|
|
|
|
col1 = Sse41.Insert(col1, mtx.MatrixCoeff00, 0);
|
|
col1 = Sse41.Insert(col1, mtx.MatrixCoeff10, 1);
|
|
col1 = Sse41.Insert(col1, mtx.MatrixCoeff20, 2);
|
|
col1 = Sse41.Insert(col1, 0, 3);
|
|
|
|
col2 = Sse41.Insert(col2, mtx.MatrixCoeff01, 0);
|
|
col2 = Sse41.Insert(col2, mtx.MatrixCoeff11, 1);
|
|
col2 = Sse41.Insert(col2, mtx.MatrixCoeff21, 2);
|
|
col2 = Sse41.Insert(col2, 0, 3);
|
|
|
|
col3 = Sse41.Insert(col3, mtx.MatrixCoeff02, 0);
|
|
col3 = Sse41.Insert(col3, mtx.MatrixCoeff12, 1);
|
|
col3 = Sse41.Insert(col3, mtx.MatrixCoeff22, 2);
|
|
col3 = Sse41.Insert(col3, one, 3);
|
|
|
|
col4 = Sse41.Insert(col4, mtx.MatrixCoeff03, 0);
|
|
col4 = Sse41.Insert(col4, mtx.MatrixCoeff13, 1);
|
|
col4 = Sse41.Insert(col4, mtx.MatrixCoeff23, 2);
|
|
col4 = Sse41.Insert(col4, 0, 3);
|
|
|
|
Vector128<int> rShift = Vector128.CreateScalar(mtx.MatrixRShift);
|
|
Vector128<ushort> clMin = Vector128.Create((ushort)slot.SlotConfig.SoftClampLow);
|
|
Vector128<ushort> clMax = Vector128.Create((ushort)slot.SlotConfig.SoftClampHigh);
|
|
|
|
fixed (Pixel* srcPtr = src.Data, dstPtr = dst.Data)
|
|
{
|
|
Pixel* ip = srcPtr;
|
|
Pixel* op = dstPtr;
|
|
|
|
for (int y = 0; y < dst.Height; y++, ip += src.Width, op += dst.Width)
|
|
{
|
|
for (int x = 0; x < dst.Width; x += 4)
|
|
{
|
|
Vector128<int> pixel1 = Sse41.ConvertToVector128Int32((ushort*)(ip + (uint)x));
|
|
Vector128<int> pixel2 = Sse41.ConvertToVector128Int32((ushort*)(ip + (uint)x + 1));
|
|
Vector128<int> pixel3 = Sse41.ConvertToVector128Int32((ushort*)(ip + (uint)x + 2));
|
|
Vector128<int> pixel4 = Sse41.ConvertToVector128Int32((ushort*)(ip + (uint)x + 3));
|
|
|
|
Vector128<ushort> pixel12, pixel34;
|
|
|
|
if (mtx.MatrixEnable)
|
|
{
|
|
pixel12 = Sse41.PackUnsignedSaturate(
|
|
MatrixMultiplySse41(pixel1, col1, col2, col3, col4, rShift),
|
|
MatrixMultiplySse41(pixel2, col1, col2, col3, col4, rShift));
|
|
pixel34 = Sse41.PackUnsignedSaturate(
|
|
MatrixMultiplySse41(pixel3, col1, col2, col3, col4, rShift),
|
|
MatrixMultiplySse41(pixel4, col1, col2, col3, col4, rShift));
|
|
}
|
|
else
|
|
{
|
|
pixel12 = Sse41.PackUnsignedSaturate(pixel1, pixel2);
|
|
pixel34 = Sse41.PackUnsignedSaturate(pixel3, pixel4);
|
|
}
|
|
|
|
pixel12 = Sse41.Min(pixel12, clMax);
|
|
pixel34 = Sse41.Min(pixel34, clMax);
|
|
pixel12 = Sse41.Max(pixel12, clMin);
|
|
pixel34 = Sse41.Max(pixel34, clMin);
|
|
|
|
Sse2.Store((ushort*)(op + (uint)x + 0), pixel12);
|
|
Sse2.Store((ushort*)(op + (uint)x + 2), pixel34);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
private static void MatrixMultiply(ref MatrixStruct mtx, int x, int y, int z, out int r, out int g, out int b)
|
|
{
|
|
if (mtx.MatrixEnable)
|
|
{
|
|
r = x * mtx.MatrixCoeff00 + y * mtx.MatrixCoeff01 + z * mtx.MatrixCoeff02;
|
|
g = x * mtx.MatrixCoeff10 + y * mtx.MatrixCoeff11 + z * mtx.MatrixCoeff12;
|
|
b = x * mtx.MatrixCoeff20 + y * mtx.MatrixCoeff21 + z * mtx.MatrixCoeff22;
|
|
|
|
r >>= mtx.MatrixRShift;
|
|
g >>= mtx.MatrixRShift;
|
|
b >>= mtx.MatrixRShift;
|
|
|
|
r += mtx.MatrixCoeff03;
|
|
g += mtx.MatrixCoeff13;
|
|
b += mtx.MatrixCoeff23;
|
|
|
|
r >>= 8;
|
|
g >>= 8;
|
|
b >>= 8;
|
|
}
|
|
else
|
|
{
|
|
r = x;
|
|
g = y;
|
|
b = z;
|
|
}
|
|
}
|
|
|
|
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
|
private static Vector128<int> MatrixMultiplySse41(
|
|
Vector128<int> pixel,
|
|
Vector128<int> col1,
|
|
Vector128<int> col2,
|
|
Vector128<int> col3,
|
|
Vector128<int> col4,
|
|
Vector128<int> rShift)
|
|
{
|
|
Vector128<int> x = Sse2.Shuffle(pixel, 0);
|
|
Vector128<int> y = Sse2.Shuffle(pixel, 0x55);
|
|
Vector128<int> z = Sse2.Shuffle(pixel, 0xea);
|
|
|
|
col1 = Sse41.MultiplyLow(col1, x);
|
|
col2 = Sse41.MultiplyLow(col2, y);
|
|
col3 = Sse41.MultiplyLow(col3, z);
|
|
|
|
Vector128<int> res = Sse2.Add(col3, Sse2.Add(col1, col2));
|
|
|
|
res = Sse2.ShiftRightArithmetic(res, rShift);
|
|
res = Sse2.Add(res, col4);
|
|
res = Sse2.ShiftRightArithmetic(res, 8);
|
|
|
|
return res;
|
|
}
|
|
}
|
|
}
|