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>
This commit is contained in:
gdkchan
GitHub
parent
26a881176e
commit
43ebd7a9bb
81 changed files with 6421 additions and 2406 deletions
|
|
@ -7,7 +7,6 @@ using Ryujinx.Graphics.Shader;
|
|||
using Ryujinx.Graphics.Texture;
|
||||
using System;
|
||||
using System.Runtime.CompilerServices;
|
||||
using System.Runtime.InteropServices;
|
||||
|
||||
namespace Ryujinx.Graphics.Gpu.Engine.Threed
|
||||
{
|
||||
|
|
@ -30,6 +29,7 @@ namespace Ryujinx.Graphics.Gpu.Engine.Threed
|
|||
private readonly StateUpdateTracker<ThreedClassState> _updateTracker;
|
||||
|
||||
private readonly ShaderProgramInfo[] _currentProgramInfo;
|
||||
private ShaderSpecializationState _shaderSpecState;
|
||||
|
||||
private bool _vtgWritesRtLayer;
|
||||
private byte _vsClipDistancesWritten;
|
||||
|
|
@ -195,6 +195,17 @@ namespace Ryujinx.Graphics.Gpu.Engine.Threed
|
|||
[MethodImpl(MethodImplOptions.AggressiveInlining)]
|
||||
public void Update()
|
||||
{
|
||||
// If any state that the shader depends on changed,
|
||||
// then we may need to compile/bind a different version
|
||||
// of the shader for the new state.
|
||||
if (_shaderSpecState != null)
|
||||
{
|
||||
if (!_shaderSpecState.MatchesGraphics(_channel, GetPoolState()))
|
||||
{
|
||||
ForceShaderUpdate();
|
||||
}
|
||||
}
|
||||
|
||||
// The vertex buffer size is calculated using a different
|
||||
// method when doing indexed draws, so we need to make sure
|
||||
// to update the vertex buffers if we are doing a regular
|
||||
|
|
@ -1065,108 +1076,127 @@ namespace Ryujinx.Graphics.Gpu.Engine.Threed
|
|||
/// </summary>
|
||||
private void UpdateShaderState()
|
||||
{
|
||||
var shaderCache = _channel.MemoryManager.Physical.ShaderCache;
|
||||
|
||||
_vtgWritesRtLayer = false;
|
||||
|
||||
ShaderAddresses addresses = new ShaderAddresses();
|
||||
|
||||
Span<ShaderAddresses> addressesSpan = MemoryMarshal.CreateSpan(ref addresses, 1);
|
||||
|
||||
Span<ulong> addressesArray = MemoryMarshal.Cast<ShaderAddresses, ulong>(addressesSpan);
|
||||
Span<ulong> addressesSpan = addresses.AsSpan();
|
||||
|
||||
ulong baseAddress = _state.State.ShaderBaseAddress.Pack();
|
||||
|
||||
for (int index = 0; index < 6; index++)
|
||||
{
|
||||
var shader = _state.State.ShaderState[index];
|
||||
|
||||
if (!shader.UnpackEnable() && index != 1)
|
||||
{
|
||||
continue;
|
||||
}
|
||||
|
||||
addressesArray[index] = baseAddress + shader.Offset;
|
||||
addressesSpan[index] = baseAddress + shader.Offset;
|
||||
}
|
||||
|
||||
GpuAccessorState gas = new GpuAccessorState(
|
||||
_state.State.TexturePoolState.Address.Pack(),
|
||||
_state.State.TexturePoolState.MaximumId,
|
||||
(int)_state.State.TextureBufferIndex,
|
||||
_state.State.EarlyZForce,
|
||||
_drawState.Topology,
|
||||
_state.State.TessMode);
|
||||
GpuChannelPoolState poolState = GetPoolState();
|
||||
GpuChannelGraphicsState graphicsState = GetGraphicsState();
|
||||
|
||||
ShaderBundle gs = _channel.MemoryManager.Physical.ShaderCache.GetGraphicsShader(ref _state.State, _channel, gas, addresses);
|
||||
CachedShaderProgram gs = shaderCache.GetGraphicsShader(ref _state.State, _channel, poolState, graphicsState, addresses);
|
||||
|
||||
_shaderSpecState = gs.SpecializationState;
|
||||
|
||||
byte oldVsClipDistancesWritten = _vsClipDistancesWritten;
|
||||
|
||||
_drawState.VsUsesInstanceId = gs.Shaders[0]?.Info.UsesInstanceId ?? false;
|
||||
_vsClipDistancesWritten = gs.Shaders[0]?.Info.ClipDistancesWritten ?? 0;
|
||||
_vtgWritesRtLayer = false;
|
||||
_drawState.VsUsesInstanceId = gs.Shaders[1]?.Info.UsesInstanceId ?? false;
|
||||
_vsClipDistancesWritten = gs.Shaders[1]?.Info.ClipDistancesWritten ?? 0;
|
||||
|
||||
if (oldVsClipDistancesWritten != _vsClipDistancesWritten)
|
||||
{
|
||||
UpdateUserClipState();
|
||||
}
|
||||
|
||||
for (int stage = 0; stage < Constants.ShaderStages; stage++)
|
||||
for (int stageIndex = 0; stageIndex < Constants.ShaderStages; stageIndex++)
|
||||
{
|
||||
ShaderProgramInfo info = gs.Shaders[stage]?.Info;
|
||||
|
||||
_currentProgramInfo[stage] = info;
|
||||
|
||||
if (info == null)
|
||||
{
|
||||
_channel.TextureManager.RentGraphicsTextureBindings(stage, 0);
|
||||
_channel.TextureManager.RentGraphicsImageBindings(stage, 0);
|
||||
_channel.BufferManager.SetGraphicsStorageBufferBindings(stage, null);
|
||||
_channel.BufferManager.SetGraphicsUniformBufferBindings(stage, null);
|
||||
continue;
|
||||
}
|
||||
|
||||
Span<TextureBindingInfo> textureBindings = _channel.TextureManager.RentGraphicsTextureBindings(stage, info.Textures.Count);
|
||||
|
||||
if (info.UsesRtLayer)
|
||||
{
|
||||
_vtgWritesRtLayer = true;
|
||||
}
|
||||
|
||||
for (int index = 0; index < info.Textures.Count; index++)
|
||||
{
|
||||
var descriptor = info.Textures[index];
|
||||
|
||||
Target target = ShaderTexture.GetTarget(descriptor.Type);
|
||||
|
||||
textureBindings[index] = new TextureBindingInfo(
|
||||
target,
|
||||
descriptor.Binding,
|
||||
descriptor.CbufSlot,
|
||||
descriptor.HandleIndex,
|
||||
descriptor.Flags);
|
||||
}
|
||||
|
||||
TextureBindingInfo[] imageBindings = _channel.TextureManager.RentGraphicsImageBindings(stage, info.Images.Count);
|
||||
|
||||
for (int index = 0; index < info.Images.Count; index++)
|
||||
{
|
||||
var descriptor = info.Images[index];
|
||||
|
||||
Target target = ShaderTexture.GetTarget(descriptor.Type);
|
||||
Format format = ShaderTexture.GetFormat(descriptor.Format);
|
||||
|
||||
imageBindings[index] = new TextureBindingInfo(
|
||||
target,
|
||||
format,
|
||||
descriptor.Binding,
|
||||
descriptor.CbufSlot,
|
||||
descriptor.HandleIndex,
|
||||
descriptor.Flags);
|
||||
}
|
||||
|
||||
_channel.BufferManager.SetGraphicsStorageBufferBindings(stage, info.SBuffers);
|
||||
_channel.BufferManager.SetGraphicsUniformBufferBindings(stage, info.CBuffers);
|
||||
UpdateStageBindings(stageIndex, gs.Shaders[stageIndex + 1]?.Info);
|
||||
}
|
||||
|
||||
_context.Renderer.Pipeline.SetProgram(gs.HostProgram);
|
||||
}
|
||||
|
||||
private void UpdateStageBindings(int stage, ShaderProgramInfo info)
|
||||
{
|
||||
_currentProgramInfo[stage] = info;
|
||||
|
||||
if (info == null)
|
||||
{
|
||||
_channel.TextureManager.RentGraphicsTextureBindings(stage, 0);
|
||||
_channel.TextureManager.RentGraphicsImageBindings(stage, 0);
|
||||
_channel.BufferManager.SetGraphicsStorageBufferBindings(stage, null);
|
||||
_channel.BufferManager.SetGraphicsUniformBufferBindings(stage, null);
|
||||
return;
|
||||
}
|
||||
|
||||
Span<TextureBindingInfo> textureBindings = _channel.TextureManager.RentGraphicsTextureBindings(stage, info.Textures.Count);
|
||||
|
||||
if (info.UsesRtLayer)
|
||||
{
|
||||
_vtgWritesRtLayer = true;
|
||||
}
|
||||
|
||||
for (int index = 0; index < info.Textures.Count; index++)
|
||||
{
|
||||
var descriptor = info.Textures[index];
|
||||
|
||||
Target target = ShaderTexture.GetTarget(descriptor.Type);
|
||||
|
||||
textureBindings[index] = new TextureBindingInfo(
|
||||
target,
|
||||
descriptor.Binding,
|
||||
descriptor.CbufSlot,
|
||||
descriptor.HandleIndex,
|
||||
descriptor.Flags);
|
||||
}
|
||||
|
||||
TextureBindingInfo[] imageBindings = _channel.TextureManager.RentGraphicsImageBindings(stage, info.Images.Count);
|
||||
|
||||
for (int index = 0; index < info.Images.Count; index++)
|
||||
{
|
||||
var descriptor = info.Images[index];
|
||||
|
||||
Target target = ShaderTexture.GetTarget(descriptor.Type);
|
||||
Format format = ShaderTexture.GetFormat(descriptor.Format);
|
||||
|
||||
imageBindings[index] = new TextureBindingInfo(
|
||||
target,
|
||||
format,
|
||||
descriptor.Binding,
|
||||
descriptor.CbufSlot,
|
||||
descriptor.HandleIndex,
|
||||
descriptor.Flags);
|
||||
}
|
||||
|
||||
_channel.BufferManager.SetGraphicsStorageBufferBindings(stage, info.SBuffers);
|
||||
_channel.BufferManager.SetGraphicsUniformBufferBindings(stage, info.CBuffers);
|
||||
}
|
||||
|
||||
private GpuChannelPoolState GetPoolState()
|
||||
{
|
||||
return new GpuChannelPoolState(
|
||||
_state.State.TexturePoolState.Address.Pack(),
|
||||
_state.State.TexturePoolState.MaximumId,
|
||||
(int)_state.State.TextureBufferIndex);
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Gets the current GPU channel state for shader creation or compatibility verification.
|
||||
/// </summary>
|
||||
/// <returns>Current GPU channel state</returns>
|
||||
private GpuChannelGraphicsState GetGraphicsState()
|
||||
{
|
||||
return new GpuChannelGraphicsState(
|
||||
_state.State.EarlyZForce,
|
||||
_drawState.Topology,
|
||||
_state.State.TessMode);
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Forces the shaders to be rebound on the next draw.
|
||||
/// </summary>
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue