ryujinx/Ryujinx.Graphics.Gpu/Shader/Cache/Migration.cs

256 lines
12 KiB
C#
Raw Normal View History

New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 15:49:44 +02:00
using Ryujinx.Common;
using Ryujinx.Common.Logging;
using Ryujinx.Common.Memory;
using Ryujinx.Graphics.GAL;
using Ryujinx.Graphics.Gpu.Engine.Threed;
using Ryujinx.Graphics.Gpu.Shader.Cache.Definition;
using Ryujinx.Graphics.Gpu.Shader.DiskCache;
using Ryujinx.Graphics.Shader;
using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.IO;
using System.Runtime.InteropServices;
namespace Ryujinx.Graphics.Gpu.Shader.Cache
{
/// <summary>
/// Class handling shader cache migrations.
/// </summary>
static class Migration
{
// Last codegen version before the migration to the new cache.
private const ulong ShaderCodeGenVersion = 3054;
/// <summary>
/// Migrates from the old cache format to the new one.
/// </summary>
/// <param name="context">GPU context</param>
/// <param name="hostStorage">Disk cache host storage (used to create the new shader files)</param>
/// <returns>Number of migrated shaders</returns>
public static int MigrateFromLegacyCache(GpuContext context, DiskCacheHostStorage hostStorage)
{
string baseCacheDirectory = CacheHelper.GetBaseCacheDirectory(GraphicsConfig.TitleId);
string cacheDirectory = CacheHelper.GenerateCachePath(baseCacheDirectory, CacheGraphicsApi.Guest, "", "program");
// If the directory does not exist, we have no old cache.
// Exist early as the CacheManager constructor will create the directories.
if (!Directory.Exists(cacheDirectory))
{
return 0;
}
if (GraphicsConfig.EnableShaderCache && GraphicsConfig.TitleId != null)
{
CacheManager cacheManager = new CacheManager(CacheGraphicsApi.OpenGL, CacheHashType.XxHash128, "glsl", GraphicsConfig.TitleId, ShaderCodeGenVersion);
bool isReadOnly = cacheManager.IsReadOnly;
HashSet<Hash128> invalidEntries = null;
if (isReadOnly)
{
Logger.Warning?.Print(LogClass.Gpu, "Loading shader cache in read-only mode (cache in use by another program!)");
}
else
{
invalidEntries = new HashSet<Hash128>();
}
ReadOnlySpan<Hash128> guestProgramList = cacheManager.GetGuestProgramList();
for (int programIndex = 0; programIndex < guestProgramList.Length; programIndex++)
{
Hash128 key = guestProgramList[programIndex];
byte[] guestProgram = cacheManager.GetGuestProgramByHash(ref key);
if (guestProgram == null)
{
Logger.Error?.Print(LogClass.Gpu, $"Ignoring orphan shader hash {key} in cache (is the cache incomplete?)");
continue;
}
ReadOnlySpan<byte> guestProgramReadOnlySpan = guestProgram;
ReadOnlySpan<GuestShaderCacheEntry> cachedShaderEntries = GuestShaderCacheEntry.Parse(ref guestProgramReadOnlySpan, out GuestShaderCacheHeader fileHeader);
if (cachedShaderEntries[0].Header.Stage == ShaderStage.Compute)
{
Debug.Assert(cachedShaderEntries.Length == 1);
GuestShaderCacheEntry entry = cachedShaderEntries[0];
byte[] code = entry.Code.AsSpan(0, entry.Header.Size - entry.Header.Cb1DataSize).ToArray();
Span<byte> codeSpan = entry.Code;
byte[] cb1Data = codeSpan.Slice(codeSpan.Length - entry.Header.Cb1DataSize).ToArray();
ShaderProgramInfo info = new ShaderProgramInfo(
Array.Empty<BufferDescriptor>(),
Array.Empty<BufferDescriptor>(),
Array.Empty<TextureDescriptor>(),
Array.Empty<TextureDescriptor>(),
ShaderStage.Compute,
false,
false,
0,
0);
GpuChannelComputeState computeState = new GpuChannelComputeState(
entry.Header.GpuAccessorHeader.ComputeLocalSizeX,
entry.Header.GpuAccessorHeader.ComputeLocalSizeY,
entry.Header.GpuAccessorHeader.ComputeLocalSizeZ,
entry.Header.GpuAccessorHeader.ComputeLocalMemorySize,
entry.Header.GpuAccessorHeader.ComputeSharedMemorySize);
ShaderSpecializationState specState = new ShaderSpecializationState(computeState);
foreach (var td in entry.TextureDescriptors)
{
var handle = td.Key;
var data = td.Value;
specState.RegisterTexture(
0,
handle,
-1,
data.UnpackFormat(),
data.UnpackSrgb(),
data.UnpackTextureTarget(),
data.UnpackTextureCoordNormalized());
}
CachedShaderStage shader = new CachedShaderStage(info, code, cb1Data);
CachedShaderProgram program = new CachedShaderProgram(null, specState, shader);
hostStorage.AddShader(context, program, ReadOnlySpan<byte>.Empty);
}
else
{
Debug.Assert(cachedShaderEntries.Length == Constants.ShaderStages);
CachedShaderStage[] shaders = new CachedShaderStage[Constants.ShaderStages + 1];
List<ShaderProgram> shaderPrograms = new List<ShaderProgram>();
TransformFeedbackDescriptorOld[] tfd = CacheHelper.ReadTransformFeedbackInformation(ref guestProgramReadOnlySpan, fileHeader);
GuestShaderCacheEntry[] entries = cachedShaderEntries.ToArray();
GuestGpuAccessorHeader accessorHeader = entries[0].Header.GpuAccessorHeader;
TessMode tessMode = new TessMode();
int tessPatchType = accessorHeader.TessellationModePacked & 3;
int tessSpacing = (accessorHeader.TessellationModePacked >> 2) & 3;
bool tessCw = (accessorHeader.TessellationModePacked & 0x10) != 0;
tessMode.Packed = (uint)tessPatchType;
tessMode.Packed |= (uint)(tessSpacing << 4);
if (tessCw)
{
tessMode.Packed |= 0x100;
}
PrimitiveTopology topology = accessorHeader.PrimitiveTopology switch
{
InputTopology.Lines => PrimitiveTopology.Lines,
InputTopology.LinesAdjacency => PrimitiveTopology.LinesAdjacency,
InputTopology.Triangles => PrimitiveTopology.Triangles,
InputTopology.TrianglesAdjacency => PrimitiveTopology.TrianglesAdjacency,
_ => PrimitiveTopology.Points
};
GpuChannelGraphicsState graphicsState = new GpuChannelGraphicsState(
accessorHeader.StateFlags.HasFlag(GuestGpuStateFlags.EarlyZForce),
topology,
tessMode,
false);
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 15:49:44 +02:00
TransformFeedbackDescriptor[] tfdNew = null;
if (tfd != null)
{
tfdNew = new TransformFeedbackDescriptor[tfd.Length];
for (int tfIndex = 0; tfIndex < tfd.Length; tfIndex++)
{
Array32<uint> varyingLocations = new Array32<uint>();
Span<byte> varyingLocationsSpan = MemoryMarshal.Cast<uint, byte>(varyingLocations.ToSpan());
tfd[tfIndex].VaryingLocations.CopyTo(varyingLocationsSpan.Slice(0, tfd[tfIndex].VaryingLocations.Length));
tfdNew[tfIndex] = new TransformFeedbackDescriptor(
tfd[tfIndex].BufferIndex,
tfd[tfIndex].Stride,
tfd[tfIndex].VaryingLocations.Length,
ref varyingLocations);
}
}
ShaderSpecializationState specState = new ShaderSpecializationState(graphicsState, tfdNew);
for (int i = 0; i < entries.Length; i++)
{
GuestShaderCacheEntry entry = entries[i];
if (entry == null)
{
continue;
}
ShaderProgramInfo info = new ShaderProgramInfo(
Array.Empty<BufferDescriptor>(),
Array.Empty<BufferDescriptor>(),
Array.Empty<TextureDescriptor>(),
Array.Empty<TextureDescriptor>(),
(ShaderStage)(i + 1),
false,
false,
0,
0);
// NOTE: Vertex B comes first in the shader cache.
byte[] code = entry.Code.AsSpan(0, entry.Header.Size - entry.Header.Cb1DataSize).ToArray();
byte[] code2 = entry.Header.SizeA != 0 ? entry.Code.AsSpan(entry.Header.Size, entry.Header.SizeA).ToArray() : null;
Span<byte> codeSpan = entry.Code;
byte[] cb1Data = codeSpan.Slice(codeSpan.Length - entry.Header.Cb1DataSize).ToArray();
shaders[i + 1] = new CachedShaderStage(info, code, cb1Data);
if (code2 != null)
{
shaders[0] = new CachedShaderStage(null, code2, cb1Data);
}
foreach (var td in entry.TextureDescriptors)
{
var handle = td.Key;
var data = td.Value;
specState.RegisterTexture(
i,
handle,
-1,
data.UnpackFormat(),
data.UnpackSrgb(),
data.UnpackTextureTarget(),
data.UnpackTextureCoordNormalized());
}
}
CachedShaderProgram program = new CachedShaderProgram(null, specState, shaders);
hostStorage.AddShader(context, program, ReadOnlySpan<byte>.Empty);
}
}
return guestProgramList.Length;
}
return 0;
}
}
}