mirror of
https://github.com/GreemDev/Ryujinx
synced 2024-12-28 22:59:46 +01:00
4d02a2d2c0
* Initial NVDEC and VIC implementation * Update FFmpeg.AutoGen to 4.3.0 * Add nvdec dependencies for Windows * Unify some VP9 structures * Rename VP9 structure fields * Improvements to Video API * XML docs for Common.Memory * Remove now unused or redundant overloads from MemoryAccessor * NVDEC UV surface read/write scalar paths * Add FIXME comments about hacky things/stuff that will need to be fixed in the future * Cleaned up VP9 memory allocation * Remove some debug logs * Rename some VP9 structs * Remove unused struct * No need to compile Ryujinx.Graphics.Host1x with unsafe anymore * Name AsyncWorkQueue threads to make debugging easier * Make Vp9PictureInfo a ref struct * LayoutConverter no longer needs the depth argument (broken by rebase) * Pooling of VP9 buffers, plus fix a memory leak on VP9 * Really wish VS could rename projects properly... * Address feedback * Remove using * Catch OperationCanceledException * Add licensing informations * Add THIRDPARTY.md to release too Co-authored-by: Thog <me@thog.eu>
133 lines
5.3 KiB
C#
133 lines
5.3 KiB
C#
using Ryujinx.Common;
|
|
using Ryujinx.Graphics.Gpu.Memory;
|
|
using Ryujinx.Graphics.Texture;
|
|
using Ryujinx.Graphics.Video;
|
|
using System;
|
|
using System.Runtime.Intrinsics;
|
|
using System.Runtime.Intrinsics.X86;
|
|
using static Ryujinx.Graphics.Nvdec.Image.SurfaceCommon;
|
|
|
|
namespace Ryujinx.Graphics.Nvdec.Image
|
|
{
|
|
static class SurfaceReader
|
|
{
|
|
public static void Read(MemoryManager gmm, ISurface surface, uint lumaOffset, uint chromaOffset)
|
|
{
|
|
int width = surface.Width;
|
|
int height = surface.Height;
|
|
int stride = surface.Stride;
|
|
|
|
ReadOnlySpan<byte> luma = gmm.DeviceGetSpan(lumaOffset, GetBlockLinearSize(width, height, 1));
|
|
|
|
ReadLuma(surface.YPlane.AsSpan(), luma, stride, width, height);
|
|
|
|
int uvWidth = surface.UvWidth;
|
|
int uvHeight = surface.UvHeight;
|
|
int uvStride = surface.UvStride;
|
|
|
|
ReadOnlySpan<byte> chroma = gmm.DeviceGetSpan(chromaOffset, GetBlockLinearSize(uvWidth, uvHeight, 2));
|
|
|
|
ReadChroma(surface.UPlane.AsSpan(), surface.VPlane.AsSpan(), chroma, uvStride, uvWidth, uvHeight);
|
|
}
|
|
|
|
private static void ReadLuma(Span<byte> dst, ReadOnlySpan<byte> src, int dstStride, int width, int height)
|
|
{
|
|
LayoutConverter.ConvertBlockLinearToLinear(dst, width, height, dstStride, 1, 2, src);
|
|
}
|
|
|
|
private unsafe static void ReadChroma(
|
|
Span<byte> dstU,
|
|
Span<byte> dstV,
|
|
ReadOnlySpan<byte> src,
|
|
int dstStride,
|
|
int width,
|
|
int height)
|
|
{
|
|
OffsetCalculator calc = new OffsetCalculator(width, height, 0, false, 2, 2);
|
|
|
|
if (Sse2.IsSupported)
|
|
{
|
|
int strideTrunc64 = BitUtils.AlignDown(width * 2, 64);
|
|
|
|
int outStrideGap = dstStride - width;
|
|
|
|
fixed (byte* dstUPtr = dstU, dstVPtr = dstV, dataPtr = src)
|
|
{
|
|
byte* uPtr = dstUPtr;
|
|
byte* vPtr = dstVPtr;
|
|
|
|
for (int y = 0; y < height; y++)
|
|
{
|
|
calc.SetY(y);
|
|
|
|
for (int x = 0; x < strideTrunc64; x += 64, uPtr += 32, vPtr += 32)
|
|
{
|
|
byte* offset = dataPtr + calc.GetOffsetWithLineOffset64(x);
|
|
byte* offset2 = offset + 0x20;
|
|
byte* offset3 = offset + 0x100;
|
|
byte* offset4 = offset + 0x120;
|
|
|
|
Vector128<byte> value = *(Vector128<byte>*)offset;
|
|
Vector128<byte> value2 = *(Vector128<byte>*)offset2;
|
|
Vector128<byte> value3 = *(Vector128<byte>*)offset3;
|
|
Vector128<byte> value4 = *(Vector128<byte>*)offset4;
|
|
|
|
Vector128<byte> u00 = Sse2.UnpackLow(value, value2);
|
|
Vector128<byte> v00 = Sse2.UnpackHigh(value, value2);
|
|
Vector128<byte> u01 = Sse2.UnpackLow(value3, value4);
|
|
Vector128<byte> v01 = Sse2.UnpackHigh(value3, value4);
|
|
|
|
Vector128<byte> u10 = Sse2.UnpackLow(u00, v00);
|
|
Vector128<byte> v10 = Sse2.UnpackHigh(u00, v00);
|
|
Vector128<byte> u11 = Sse2.UnpackLow(u01, v01);
|
|
Vector128<byte> v11 = Sse2.UnpackHigh(u01, v01);
|
|
|
|
Vector128<byte> u20 = Sse2.UnpackLow(u10, v10);
|
|
Vector128<byte> v20 = Sse2.UnpackHigh(u10, v10);
|
|
Vector128<byte> u21 = Sse2.UnpackLow(u11, v11);
|
|
Vector128<byte> v21 = Sse2.UnpackHigh(u11, v11);
|
|
|
|
Vector128<byte> u30 = Sse2.UnpackLow(u20, v20);
|
|
Vector128<byte> v30 = Sse2.UnpackHigh(u20, v20);
|
|
Vector128<byte> u31 = Sse2.UnpackLow(u21, v21);
|
|
Vector128<byte> v31 = Sse2.UnpackHigh(u21, v21);
|
|
|
|
*(Vector128<byte>*)uPtr = u30;
|
|
*(Vector128<byte>*)(uPtr + 16) = u31;
|
|
*(Vector128<byte>*)vPtr = v30;
|
|
*(Vector128<byte>*)(vPtr + 16) = v31;
|
|
}
|
|
|
|
for (int x = strideTrunc64 / 2; x < width; x++, uPtr++, vPtr++)
|
|
{
|
|
byte* offset = dataPtr + calc.GetOffset(x);
|
|
|
|
*uPtr = *offset;
|
|
*vPtr = *(offset + 1);
|
|
}
|
|
|
|
uPtr += outStrideGap;
|
|
vPtr += outStrideGap;
|
|
}
|
|
}
|
|
}
|
|
else
|
|
{
|
|
for (int y = 0; y < height; y++)
|
|
{
|
|
int dstBaseOffset = y * dstStride;
|
|
|
|
calc.SetY(y);
|
|
|
|
for (int x = 0; x < width; x++)
|
|
{
|
|
int srcOffset = calc.GetOffset(x);
|
|
|
|
dstU[dstBaseOffset + x] = src[srcOffset];
|
|
dstV[dstBaseOffset + x] = src[srcOffset + 1];
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|