Ryujinx-uplift/src/Ryujinx.Graphics.GAL/Multithreading/CommandHelper.cs

151 lines
8.7 KiB
C#
Raw Normal View History

using Ryujinx.Graphics.GAL.Multithreading.Commands;
Add a Multithreading layer for the GAL, multi-thread shader compilation at runtime (#2501) * Initial Implementation About as fast as nvidia GL multithreading, can be improved with faster command queuing. * Struct based command list Speeds up a bit. Still a lot of time lost to resource copy. * Do shader init while the render thread is active. * Introduce circular span pool V1 Ideally should be able to use structs instead of references for storing these spans on commands. Will try that next. * Refactor SpanRef some more Use a struct to represent SpanRef, rather than a reference. * Flush buffers on background thread * Use a span for UpdateRenderScale. Much faster than copying the array. * Calculate command size using reflection * WIP parallel shaders * Some minor optimisation * Only 2 max refs per command now. The command with 3 refs is gone. :relieved: * Don't cast on the GPU side * Remove redundant casts, force sync on window present * Fix Shader Cache * Fix host shader save. * Fixup to work with new renderer stuff * Make command Run static, use array of delegates as lookup Profile says this takes less time than the previous way. * Bring up to date * Add settings toggle. Fix Muiltithreading Off mode. * Fix warning. * Release tracking lock for flushes * Fix Conditional Render fast path with threaded gal * Make handle iteration safe when releasing the lock This is mostly temporary. * Attempt to set backend threading on driver Only really works on nvidia before launching a game. * Fix race condition with BufferModifiedRangeList, exceptions in tracking actions * Update buffer set commands * Some cleanup * Only use stutter workaround when using opengl renderer non-threaded * Add host-conditional reservation of counter events There has always been the possibility that conditional rendering could use a query object just as it is disposed by the counter queue. This change makes it so that when the host decides to use host conditional rendering, the query object is reserved so that it cannot be deleted. Counter events can optionally start reserved, as the threaded implementation can reserve them before the backend creates them, and there would otherwise be a short amount of time where the counter queue could dispose the event before a call to reserve it could be made. * Address Feedback * Make counter flush tracked again. Hopefully does not cause any issues this time. * Wait for FlushTo on the main queue thread. Currently assumes only one thread will want to FlushTo (in this case, the GPU thread) * Add SDL2 headless integration * Add HLE macro commands. Co-authored-by: Mary <mary@mary.zone>
2021-08-27 00:31:29 +02:00
using Ryujinx.Graphics.GAL.Multithreading.Commands.Buffer;
using Ryujinx.Graphics.GAL.Multithreading.Commands.CounterEvent;
using Ryujinx.Graphics.GAL.Multithreading.Commands.Program;
using Ryujinx.Graphics.GAL.Multithreading.Commands.Renderer;
using Ryujinx.Graphics.GAL.Multithreading.Commands.Sampler;
using Ryujinx.Graphics.GAL.Multithreading.Commands.Texture;
using Ryujinx.Graphics.GAL.Multithreading.Commands.Window;
using System;
using System.Linq;
using System.Runtime.CompilerServices;
using System.Runtime.InteropServices;
namespace Ryujinx.Graphics.GAL.Multithreading
{
static class CommandHelper
{
private delegate void CommandDelegate(Span<byte> memory, ThreadedRenderer threaded, IRenderer renderer);
[Ryujinx.Graphics.GAL] Address dotnet-format issues (#5366) * dotnet format style --severity info Some changes were manually reverted. * dotnet format analyzers --serverity info Some changes have been minimally adapted. * Restore a few unused methods and variables * Silence dotnet format IDE0052 warnings * Address dotnet format CA1816 warnings * Address or silence dotnet format CA1069 warnings * Address remaining dotnet format analyzer warnings * Address review comments * Address most dotnet format whitespace warnings * Apply dotnet format whitespace formatting A few of them have been manually reverted and the corresponding warning was silenced * Revert formatting changes for while and for-loops * Another rebase, another dotnet format run * Run dotnet format whitespace after rebase * Run dotnet format style after rebase * Run dotnet format analyzers after rebase * Run dotnet format after rebase and remove unused usings - analyzers - style - whitespace * Disable 'prefer switch expression' rule * Add comments to disabled warnings * Simplify properties and array initialization, Use const when possible, Remove trailing commas * Start working on disabled warnings * Address IDE0251 warnings * Revert "Simplify properties and array initialization, Use const when possible, Remove trailing commas" This reverts commit 9462e4136c0a2100dc28b20cf9542e06790aa67e. * dotnet format whitespace after rebase * First dotnet format pass * Address review feedback * Add trailing commas * Remove SuppressMessage for IDE0066 * Make explicit Equals implementation implicit
2023-06-28 20:20:10 +02:00
private static readonly int _totalCommands = (int)Enum.GetValues<CommandType>().Max() + 1;
private static readonly CommandDelegate[] _lookup = new CommandDelegate[_totalCommands];
Add a Multithreading layer for the GAL, multi-thread shader compilation at runtime (#2501) * Initial Implementation About as fast as nvidia GL multithreading, can be improved with faster command queuing. * Struct based command list Speeds up a bit. Still a lot of time lost to resource copy. * Do shader init while the render thread is active. * Introduce circular span pool V1 Ideally should be able to use structs instead of references for storing these spans on commands. Will try that next. * Refactor SpanRef some more Use a struct to represent SpanRef, rather than a reference. * Flush buffers on background thread * Use a span for UpdateRenderScale. Much faster than copying the array. * Calculate command size using reflection * WIP parallel shaders * Some minor optimisation * Only 2 max refs per command now. The command with 3 refs is gone. :relieved: * Don't cast on the GPU side * Remove redundant casts, force sync on window present * Fix Shader Cache * Fix host shader save. * Fixup to work with new renderer stuff * Make command Run static, use array of delegates as lookup Profile says this takes less time than the previous way. * Bring up to date * Add settings toggle. Fix Muiltithreading Off mode. * Fix warning. * Release tracking lock for flushes * Fix Conditional Render fast path with threaded gal * Make handle iteration safe when releasing the lock This is mostly temporary. * Attempt to set backend threading on driver Only really works on nvidia before launching a game. * Fix race condition with BufferModifiedRangeList, exceptions in tracking actions * Update buffer set commands * Some cleanup * Only use stutter workaround when using opengl renderer non-threaded * Add host-conditional reservation of counter events There has always been the possibility that conditional rendering could use a query object just as it is disposed by the counter queue. This change makes it so that when the host decides to use host conditional rendering, the query object is reserved so that it cannot be deleted. Counter events can optionally start reserved, as the threaded implementation can reserve them before the backend creates them, and there would otherwise be a short amount of time where the counter queue could dispose the event before a call to reserve it could be made. * Address Feedback * Make counter flush tracked again. Hopefully does not cause any issues this time. * Wait for FlushTo on the main queue thread. Currently assumes only one thread will want to FlushTo (in this case, the GPU thread) * Add SDL2 headless integration * Add HLE macro commands. Co-authored-by: Mary <mary@mary.zone>
2021-08-27 00:31:29 +02:00
[MethodImpl(MethodImplOptions.AggressiveInlining)]
private static ref T GetCommand<T>(Span<byte> memory)
{
return ref Unsafe.As<byte, T>(ref MemoryMarshal.GetReference(memory));
}
public static int GetMaxCommandSize()
{
return InitLookup() + 1; // 1 byte reserved for command size.
Add a Multithreading layer for the GAL, multi-thread shader compilation at runtime (#2501) * Initial Implementation About as fast as nvidia GL multithreading, can be improved with faster command queuing. * Struct based command list Speeds up a bit. Still a lot of time lost to resource copy. * Do shader init while the render thread is active. * Introduce circular span pool V1 Ideally should be able to use structs instead of references for storing these spans on commands. Will try that next. * Refactor SpanRef some more Use a struct to represent SpanRef, rather than a reference. * Flush buffers on background thread * Use a span for UpdateRenderScale. Much faster than copying the array. * Calculate command size using reflection * WIP parallel shaders * Some minor optimisation * Only 2 max refs per command now. The command with 3 refs is gone. :relieved: * Don't cast on the GPU side * Remove redundant casts, force sync on window present * Fix Shader Cache * Fix host shader save. * Fixup to work with new renderer stuff * Make command Run static, use array of delegates as lookup Profile says this takes less time than the previous way. * Bring up to date * Add settings toggle. Fix Muiltithreading Off mode. * Fix warning. * Release tracking lock for flushes * Fix Conditional Render fast path with threaded gal * Make handle iteration safe when releasing the lock This is mostly temporary. * Attempt to set backend threading on driver Only really works on nvidia before launching a game. * Fix race condition with BufferModifiedRangeList, exceptions in tracking actions * Update buffer set commands * Some cleanup * Only use stutter workaround when using opengl renderer non-threaded * Add host-conditional reservation of counter events There has always been the possibility that conditional rendering could use a query object just as it is disposed by the counter queue. This change makes it so that when the host decides to use host conditional rendering, the query object is reserved so that it cannot be deleted. Counter events can optionally start reserved, as the threaded implementation can reserve them before the backend creates them, and there would otherwise be a short amount of time where the counter queue could dispose the event before a call to reserve it could be made. * Address Feedback * Make counter flush tracked again. Hopefully does not cause any issues this time. * Wait for FlushTo on the main queue thread. Currently assumes only one thread will want to FlushTo (in this case, the GPU thread) * Add SDL2 headless integration * Add HLE macro commands. Co-authored-by: Mary <mary@mary.zone>
2021-08-27 00:31:29 +02:00
}
private static int InitLookup()
Add a Multithreading layer for the GAL, multi-thread shader compilation at runtime (#2501) * Initial Implementation About as fast as nvidia GL multithreading, can be improved with faster command queuing. * Struct based command list Speeds up a bit. Still a lot of time lost to resource copy. * Do shader init while the render thread is active. * Introduce circular span pool V1 Ideally should be able to use structs instead of references for storing these spans on commands. Will try that next. * Refactor SpanRef some more Use a struct to represent SpanRef, rather than a reference. * Flush buffers on background thread * Use a span for UpdateRenderScale. Much faster than copying the array. * Calculate command size using reflection * WIP parallel shaders * Some minor optimisation * Only 2 max refs per command now. The command with 3 refs is gone. :relieved: * Don't cast on the GPU side * Remove redundant casts, force sync on window present * Fix Shader Cache * Fix host shader save. * Fixup to work with new renderer stuff * Make command Run static, use array of delegates as lookup Profile says this takes less time than the previous way. * Bring up to date * Add settings toggle. Fix Muiltithreading Off mode. * Fix warning. * Release tracking lock for flushes * Fix Conditional Render fast path with threaded gal * Make handle iteration safe when releasing the lock This is mostly temporary. * Attempt to set backend threading on driver Only really works on nvidia before launching a game. * Fix race condition with BufferModifiedRangeList, exceptions in tracking actions * Update buffer set commands * Some cleanup * Only use stutter workaround when using opengl renderer non-threaded * Add host-conditional reservation of counter events There has always been the possibility that conditional rendering could use a query object just as it is disposed by the counter queue. This change makes it so that when the host decides to use host conditional rendering, the query object is reserved so that it cannot be deleted. Counter events can optionally start reserved, as the threaded implementation can reserve them before the backend creates them, and there would otherwise be a short amount of time where the counter queue could dispose the event before a call to reserve it could be made. * Address Feedback * Make counter flush tracked again. Hopefully does not cause any issues this time. * Wait for FlushTo on the main queue thread. Currently assumes only one thread will want to FlushTo (in this case, the GPU thread) * Add SDL2 headless integration * Add HLE macro commands. Co-authored-by: Mary <mary@mary.zone>
2021-08-27 00:31:29 +02:00
{
int maxCommandSize = 0;
void Register<T>(CommandType commandType) where T : unmanaged, IGALCommand, IGALCommand<T>
{
maxCommandSize = Math.Max(maxCommandSize, Unsafe.SizeOf<T>());
_lookup[(int)commandType] = (memory, threaded, renderer) => T.Run(ref GetCommand<T>(memory), threaded, renderer);
}
Register<ActionCommand>(CommandType.Action);
Register<CreateBufferCommand>(CommandType.CreateBuffer);
Register<CreateBufferAccessCommand>(CommandType.CreateBufferAccess);
Register<CreateHostBufferCommand>(CommandType.CreateHostBuffer);
Register<CreateProgramCommand>(CommandType.CreateProgram);
Register<CreateSamplerCommand>(CommandType.CreateSampler);
Register<CreateSyncCommand>(CommandType.CreateSync);
Register<CreateTextureCommand>(CommandType.CreateTexture);
Register<GetCapabilitiesCommand>(CommandType.GetCapabilities);
Register<PreFrameCommand>(CommandType.PreFrame);
Register<ReportCounterCommand>(CommandType.ReportCounter);
Register<ResetCounterCommand>(CommandType.ResetCounter);
Register<UpdateCountersCommand>(CommandType.UpdateCounters);
Register<BufferDisposeCommand>(CommandType.BufferDispose);
Register<BufferGetDataCommand>(CommandType.BufferGetData);
Register<BufferSetDataCommand>(CommandType.BufferSetData);
Register<CounterEventDisposeCommand>(CommandType.CounterEventDispose);
Register<CounterEventFlushCommand>(CommandType.CounterEventFlush);
Register<ProgramDisposeCommand>(CommandType.ProgramDispose);
Register<ProgramGetBinaryCommand>(CommandType.ProgramGetBinary);
Register<ProgramCheckLinkCommand>(CommandType.ProgramCheckLink);
Register<SamplerDisposeCommand>(CommandType.SamplerDispose);
Register<TextureCopyToCommand>(CommandType.TextureCopyTo);
Register<TextureCopyToScaledCommand>(CommandType.TextureCopyToScaled);
Register<TextureCopyToSliceCommand>(CommandType.TextureCopyToSlice);
Register<TextureCopyToBufferCommand>(CommandType.TextureCopyToBuffer);
Register<TextureCreateViewCommand>(CommandType.TextureCreateView);
Register<TextureGetDataCommand>(CommandType.TextureGetData);
Register<TextureGetDataSliceCommand>(CommandType.TextureGetDataSlice);
Register<TextureReleaseCommand>(CommandType.TextureRelease);
Register<TextureSetDataCommand>(CommandType.TextureSetData);
Register<TextureSetDataSliceCommand>(CommandType.TextureSetDataSlice);
Register<TextureSetDataSliceRegionCommand>(CommandType.TextureSetDataSliceRegion);
Register<TextureSetStorageCommand>(CommandType.TextureSetStorage);
Register<WindowPresentCommand>(CommandType.WindowPresent);
Register<BarrierCommand>(CommandType.Barrier);
Register<BeginTransformFeedbackCommand>(CommandType.BeginTransformFeedback);
Register<ClearBufferCommand>(CommandType.ClearBuffer);
Register<ClearRenderTargetColorCommand>(CommandType.ClearRenderTargetColor);
Register<ClearRenderTargetDepthStencilCommand>(CommandType.ClearRenderTargetDepthStencil);
Register<CommandBufferBarrierCommand>(CommandType.CommandBufferBarrier);
Register<CopyBufferCommand>(CommandType.CopyBuffer);
Register<DispatchComputeCommand>(CommandType.DispatchCompute);
Register<DrawCommand>(CommandType.Draw);
Register<DrawIndexedCommand>(CommandType.DrawIndexed);
Register<DrawIndexedIndirectCommand>(CommandType.DrawIndexedIndirect);
Register<DrawIndexedIndirectCountCommand>(CommandType.DrawIndexedIndirectCount);
Register<DrawIndirectCommand>(CommandType.DrawIndirect);
Register<DrawIndirectCountCommand>(CommandType.DrawIndirectCount);
Register<DrawTextureCommand>(CommandType.DrawTexture);
Register<EndHostConditionalRenderingCommand>(CommandType.EndHostConditionalRendering);
Register<EndTransformFeedbackCommand>(CommandType.EndTransformFeedback);
Register<SetAlphaTestCommand>(CommandType.SetAlphaTest);
Register<SetBlendStateAdvancedCommand>(CommandType.SetBlendStateAdvanced);
Register<SetBlendStateCommand>(CommandType.SetBlendState);
Register<SetDepthBiasCommand>(CommandType.SetDepthBias);
Register<SetDepthClampCommand>(CommandType.SetDepthClamp);
Register<SetDepthModeCommand>(CommandType.SetDepthMode);
Register<SetDepthTestCommand>(CommandType.SetDepthTest);
Register<SetFaceCullingCommand>(CommandType.SetFaceCulling);
Register<SetFrontFaceCommand>(CommandType.SetFrontFace);
Register<SetStorageBuffersCommand>(CommandType.SetStorageBuffers);
Register<SetTransformFeedbackBuffersCommand>(CommandType.SetTransformFeedbackBuffers);
Register<SetUniformBuffersCommand>(CommandType.SetUniformBuffers);
Register<SetImageCommand>(CommandType.SetImage);
Register<SetIndexBufferCommand>(CommandType.SetIndexBuffer);
Register<SetLineParametersCommand>(CommandType.SetLineParameters);
Register<SetLogicOpStateCommand>(CommandType.SetLogicOpState);
Register<SetMultisampleStateCommand>(CommandType.SetMultisampleState);
Register<SetPatchParametersCommand>(CommandType.SetPatchParameters);
Register<SetPointParametersCommand>(CommandType.SetPointParameters);
Register<SetPolygonModeCommand>(CommandType.SetPolygonMode);
Register<SetPrimitiveRestartCommand>(CommandType.SetPrimitiveRestart);
Register<SetPrimitiveTopologyCommand>(CommandType.SetPrimitiveTopology);
Register<SetProgramCommand>(CommandType.SetProgram);
Register<SetRasterizerDiscardCommand>(CommandType.SetRasterizerDiscard);
Register<SetRenderTargetColorMasksCommand>(CommandType.SetRenderTargetColorMasks);
Register<SetRenderTargetsCommand>(CommandType.SetRenderTargets);
Register<SetScissorsCommand>(CommandType.SetScissor);
Register<SetStencilTestCommand>(CommandType.SetStencilTest);
Register<SetTextureAndSamplerCommand>(CommandType.SetTextureAndSampler);
Register<SetUserClipDistanceCommand>(CommandType.SetUserClipDistance);
Register<SetVertexAttribsCommand>(CommandType.SetVertexAttribs);
Register<SetVertexBuffersCommand>(CommandType.SetVertexBuffers);
Register<SetViewportsCommand>(CommandType.SetViewports);
Register<TextureBarrierCommand>(CommandType.TextureBarrier);
Register<TextureBarrierTiledCommand>(CommandType.TextureBarrierTiled);
Register<TryHostConditionalRenderingCommand>(CommandType.TryHostConditionalRendering);
Register<TryHostConditionalRenderingFlushCommand>(CommandType.TryHostConditionalRenderingFlush);
return maxCommandSize;
Add a Multithreading layer for the GAL, multi-thread shader compilation at runtime (#2501) * Initial Implementation About as fast as nvidia GL multithreading, can be improved with faster command queuing. * Struct based command list Speeds up a bit. Still a lot of time lost to resource copy. * Do shader init while the render thread is active. * Introduce circular span pool V1 Ideally should be able to use structs instead of references for storing these spans on commands. Will try that next. * Refactor SpanRef some more Use a struct to represent SpanRef, rather than a reference. * Flush buffers on background thread * Use a span for UpdateRenderScale. Much faster than copying the array. * Calculate command size using reflection * WIP parallel shaders * Some minor optimisation * Only 2 max refs per command now. The command with 3 refs is gone. :relieved: * Don't cast on the GPU side * Remove redundant casts, force sync on window present * Fix Shader Cache * Fix host shader save. * Fixup to work with new renderer stuff * Make command Run static, use array of delegates as lookup Profile says this takes less time than the previous way. * Bring up to date * Add settings toggle. Fix Muiltithreading Off mode. * Fix warning. * Release tracking lock for flushes * Fix Conditional Render fast path with threaded gal * Make handle iteration safe when releasing the lock This is mostly temporary. * Attempt to set backend threading on driver Only really works on nvidia before launching a game. * Fix race condition with BufferModifiedRangeList, exceptions in tracking actions * Update buffer set commands * Some cleanup * Only use stutter workaround when using opengl renderer non-threaded * Add host-conditional reservation of counter events There has always been the possibility that conditional rendering could use a query object just as it is disposed by the counter queue. This change makes it so that when the host decides to use host conditional rendering, the query object is reserved so that it cannot be deleted. Counter events can optionally start reserved, as the threaded implementation can reserve them before the backend creates them, and there would otherwise be a short amount of time where the counter queue could dispose the event before a call to reserve it could be made. * Address Feedback * Make counter flush tracked again. Hopefully does not cause any issues this time. * Wait for FlushTo on the main queue thread. Currently assumes only one thread will want to FlushTo (in this case, the GPU thread) * Add SDL2 headless integration * Add HLE macro commands. Co-authored-by: Mary <mary@mary.zone>
2021-08-27 00:31:29 +02:00
}
[MethodImpl(MethodImplOptions.AggressiveInlining)]
public static void RunCommand(Span<byte> memory, ThreadedRenderer threaded, IRenderer renderer)
{
[Ryujinx.Graphics.GAL] Address dotnet-format issues (#5366) * dotnet format style --severity info Some changes were manually reverted. * dotnet format analyzers --serverity info Some changes have been minimally adapted. * Restore a few unused methods and variables * Silence dotnet format IDE0052 warnings * Address dotnet format CA1816 warnings * Address or silence dotnet format CA1069 warnings * Address remaining dotnet format analyzer warnings * Address review comments * Address most dotnet format whitespace warnings * Apply dotnet format whitespace formatting A few of them have been manually reverted and the corresponding warning was silenced * Revert formatting changes for while and for-loops * Another rebase, another dotnet format run * Run dotnet format whitespace after rebase * Run dotnet format style after rebase * Run dotnet format analyzers after rebase * Run dotnet format after rebase and remove unused usings - analyzers - style - whitespace * Disable 'prefer switch expression' rule * Add comments to disabled warnings * Simplify properties and array initialization, Use const when possible, Remove trailing commas * Start working on disabled warnings * Address IDE0251 warnings * Revert "Simplify properties and array initialization, Use const when possible, Remove trailing commas" This reverts commit 9462e4136c0a2100dc28b20cf9542e06790aa67e. * dotnet format whitespace after rebase * First dotnet format pass * Address review feedback * Add trailing commas * Remove SuppressMessage for IDE0066 * Make explicit Equals implementation implicit
2023-06-28 20:20:10 +02:00
_lookup[memory[^1]](memory, threaded, renderer);
Add a Multithreading layer for the GAL, multi-thread shader compilation at runtime (#2501) * Initial Implementation About as fast as nvidia GL multithreading, can be improved with faster command queuing. * Struct based command list Speeds up a bit. Still a lot of time lost to resource copy. * Do shader init while the render thread is active. * Introduce circular span pool V1 Ideally should be able to use structs instead of references for storing these spans on commands. Will try that next. * Refactor SpanRef some more Use a struct to represent SpanRef, rather than a reference. * Flush buffers on background thread * Use a span for UpdateRenderScale. Much faster than copying the array. * Calculate command size using reflection * WIP parallel shaders * Some minor optimisation * Only 2 max refs per command now. The command with 3 refs is gone. :relieved: * Don't cast on the GPU side * Remove redundant casts, force sync on window present * Fix Shader Cache * Fix host shader save. * Fixup to work with new renderer stuff * Make command Run static, use array of delegates as lookup Profile says this takes less time than the previous way. * Bring up to date * Add settings toggle. Fix Muiltithreading Off mode. * Fix warning. * Release tracking lock for flushes * Fix Conditional Render fast path with threaded gal * Make handle iteration safe when releasing the lock This is mostly temporary. * Attempt to set backend threading on driver Only really works on nvidia before launching a game. * Fix race condition with BufferModifiedRangeList, exceptions in tracking actions * Update buffer set commands * Some cleanup * Only use stutter workaround when using opengl renderer non-threaded * Add host-conditional reservation of counter events There has always been the possibility that conditional rendering could use a query object just as it is disposed by the counter queue. This change makes it so that when the host decides to use host conditional rendering, the query object is reserved so that it cannot be deleted. Counter events can optionally start reserved, as the threaded implementation can reserve them before the backend creates them, and there would otherwise be a short amount of time where the counter queue could dispose the event before a call to reserve it could be made. * Address Feedback * Make counter flush tracked again. Hopefully does not cause any issues this time. * Wait for FlushTo on the main queue thread. Currently assumes only one thread will want to FlushTo (in this case, the GPU thread) * Add SDL2 headless integration * Add HLE macro commands. Co-authored-by: Mary <mary@mary.zone>
2021-08-27 00:31:29 +02:00
}
}
}