GPU: Pass SpanOrArray for Texture SetData to avoid copy (#3745)
* GPU: Pass SpanOrArray for Texture SetData to avoid copy
Texture data is often converted before upload, meaning that an array was allocated to perform the conversion into. However, the backend SetData methods were being passed a Span of that data, and the Multithreaded layer does `ToArray()` on it so that it can be stored for later! This method can't extract the original array, so it creates a copy.
This PR changes the type passed for textures to a new ref struct called SpanOrArray, which is backed by either a ReadOnlySpan or an array. The benefit here is that we can have a ToArray method that doesn't copy if it is originally backed by an array.
This will also avoid a copy when running the ASTC decoder.
On NieR this was taking 38% of texture upload time, which it does a _lot_ of when you move between areas, so there should be a 1.6x performance boost when strictly uploading textures. No doubt this will also improve texture streaming performance in UE4 games, and maybe a small reduction with video playback.
From the numbers, it's probably possible to improve the upload rate by a further 1.6x by performing layout conversion on GPU. I'm not sure if we could improve it further than that - multithreading conversion on CPU would probably result in memory bottleneck.
This doesn't extend to buffers, since we don't convert their data on the GPU emulator side.
* Remove implicit cast to array.
2022-10-08 16:04:47 +01:00
|
|
|
|
using Ryujinx.Common.Memory;
|
|
|
|
|
using Ryujinx.Graphics.GAL.Multithreading.Commands.Texture;
|
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>
2021-08-26 23:31:29 +01:00
|
|
|
|
using Ryujinx.Graphics.GAL.Multithreading.Model;
|
|
|
|
|
using System;
|
|
|
|
|
|
|
|
|
|
namespace Ryujinx.Graphics.GAL.Multithreading.Resources
|
|
|
|
|
{
|
|
|
|
|
/// <summary>
|
|
|
|
|
/// Threaded representation of a texture.
|
|
|
|
|
/// </summary>
|
|
|
|
|
class ThreadedTexture : ITexture
|
|
|
|
|
{
|
|
|
|
|
private ThreadedRenderer _renderer;
|
|
|
|
|
private TextureCreateInfo _info;
|
|
|
|
|
public ITexture Base;
|
|
|
|
|
|
|
|
|
|
public int Width => _info.Width;
|
|
|
|
|
|
|
|
|
|
public int Height => _info.Height;
|
|
|
|
|
|
|
|
|
|
public float ScaleFactor { get; }
|
|
|
|
|
|
|
|
|
|
public ThreadedTexture(ThreadedRenderer renderer, TextureCreateInfo info, float scale)
|
|
|
|
|
{
|
|
|
|
|
_renderer = renderer;
|
|
|
|
|
_info = info;
|
|
|
|
|
ScaleFactor = scale;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
private TableRef<T> Ref<T>(T reference)
|
|
|
|
|
{
|
|
|
|
|
return new TableRef<T>(_renderer, reference);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public void CopyTo(ITexture destination, int firstLayer, int firstLevel)
|
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureCopyToCommand>().Set(Ref(this), Ref((ThreadedTexture)destination), firstLayer, firstLevel);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public void CopyTo(ITexture destination, int srcLayer, int dstLayer, int srcLevel, int dstLevel)
|
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureCopyToSliceCommand>().Set(Ref(this), Ref((ThreadedTexture)destination), srcLayer, dstLayer, srcLevel, dstLevel);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public void CopyTo(ITexture destination, Extents2D srcRegion, Extents2D dstRegion, bool linearFilter)
|
|
|
|
|
{
|
|
|
|
|
ThreadedTexture dest = (ThreadedTexture)destination;
|
|
|
|
|
|
|
|
|
|
if (_renderer.IsGpuThread())
|
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureCopyToScaledCommand>().Set(Ref(this), Ref(dest), srcRegion, dstRegion, linearFilter);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
// Scaled copy can happen on another thread for a res scale flush.
|
|
|
|
|
ThreadedHelpers.SpinUntilNonNull(ref Base);
|
|
|
|
|
ThreadedHelpers.SpinUntilNonNull(ref dest.Base);
|
|
|
|
|
|
|
|
|
|
Base.CopyTo(dest.Base, srcRegion, dstRegion, linearFilter);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public ITexture CreateView(TextureCreateInfo info, int firstLayer, int firstLevel)
|
|
|
|
|
{
|
|
|
|
|
ThreadedTexture newTex = new ThreadedTexture(_renderer, info, ScaleFactor);
|
|
|
|
|
_renderer.New<TextureCreateViewCommand>().Set(Ref(this), Ref(newTex), info, firstLayer, firstLevel);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
|
|
|
|
|
return newTex;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public ReadOnlySpan<byte> GetData()
|
|
|
|
|
{
|
|
|
|
|
if (_renderer.IsGpuThread())
|
|
|
|
|
{
|
|
|
|
|
ResultBox<PinnedSpan<byte>> box = new ResultBox<PinnedSpan<byte>>();
|
|
|
|
|
_renderer.New<TextureGetDataCommand>().Set(Ref(this), Ref(box));
|
|
|
|
|
_renderer.InvokeCommand();
|
|
|
|
|
|
|
|
|
|
return box.Result.Get();
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
ThreadedHelpers.SpinUntilNonNull(ref Base);
|
|
|
|
|
|
|
|
|
|
return Base.GetData();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2022-01-09 16:28:48 +00:00
|
|
|
|
public ReadOnlySpan<byte> GetData(int layer, int level)
|
|
|
|
|
{
|
|
|
|
|
if (_renderer.IsGpuThread())
|
|
|
|
|
{
|
|
|
|
|
ResultBox<PinnedSpan<byte>> box = new ResultBox<PinnedSpan<byte>>();
|
|
|
|
|
_renderer.New<TextureGetDataSliceCommand>().Set(Ref(this), Ref(box), layer, level);
|
|
|
|
|
_renderer.InvokeCommand();
|
|
|
|
|
|
|
|
|
|
return box.Result.Get();
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
ThreadedHelpers.SpinUntilNonNull(ref Base);
|
|
|
|
|
|
|
|
|
|
return Base.GetData(layer, level);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
GPU: Pass SpanOrArray for Texture SetData to avoid copy (#3745)
* GPU: Pass SpanOrArray for Texture SetData to avoid copy
Texture data is often converted before upload, meaning that an array was allocated to perform the conversion into. However, the backend SetData methods were being passed a Span of that data, and the Multithreaded layer does `ToArray()` on it so that it can be stored for later! This method can't extract the original array, so it creates a copy.
This PR changes the type passed for textures to a new ref struct called SpanOrArray, which is backed by either a ReadOnlySpan or an array. The benefit here is that we can have a ToArray method that doesn't copy if it is originally backed by an array.
This will also avoid a copy when running the ASTC decoder.
On NieR this was taking 38% of texture upload time, which it does a _lot_ of when you move between areas, so there should be a 1.6x performance boost when strictly uploading textures. No doubt this will also improve texture streaming performance in UE4 games, and maybe a small reduction with video playback.
From the numbers, it's probably possible to improve the upload rate by a further 1.6x by performing layout conversion on GPU. I'm not sure if we could improve it further than that - multithreading conversion on CPU would probably result in memory bottleneck.
This doesn't extend to buffers, since we don't convert their data on the GPU emulator side.
* Remove implicit cast to array.
2022-10-08 16:04:47 +01:00
|
|
|
|
public void SetData(SpanOrArray<byte> data)
|
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>
2021-08-26 23:31:29 +01:00
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureSetDataCommand>().Set(Ref(this), Ref(data.ToArray()));
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
|
GPU: Pass SpanOrArray for Texture SetData to avoid copy (#3745)
* GPU: Pass SpanOrArray for Texture SetData to avoid copy
Texture data is often converted before upload, meaning that an array was allocated to perform the conversion into. However, the backend SetData methods were being passed a Span of that data, and the Multithreaded layer does `ToArray()` on it so that it can be stored for later! This method can't extract the original array, so it creates a copy.
This PR changes the type passed for textures to a new ref struct called SpanOrArray, which is backed by either a ReadOnlySpan or an array. The benefit here is that we can have a ToArray method that doesn't copy if it is originally backed by an array.
This will also avoid a copy when running the ASTC decoder.
On NieR this was taking 38% of texture upload time, which it does a _lot_ of when you move between areas, so there should be a 1.6x performance boost when strictly uploading textures. No doubt this will also improve texture streaming performance in UE4 games, and maybe a small reduction with video playback.
From the numbers, it's probably possible to improve the upload rate by a further 1.6x by performing layout conversion on GPU. I'm not sure if we could improve it further than that - multithreading conversion on CPU would probably result in memory bottleneck.
This doesn't extend to buffers, since we don't convert their data on the GPU emulator side.
* Remove implicit cast to array.
2022-10-08 16:04:47 +01:00
|
|
|
|
public void SetData(SpanOrArray<byte> data, int layer, int level)
|
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>
2021-08-26 23:31:29 +01:00
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureSetDataSliceCommand>().Set(Ref(this), Ref(data.ToArray()), layer, level);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
|
GPU: Pass SpanOrArray for Texture SetData to avoid copy (#3745)
* GPU: Pass SpanOrArray for Texture SetData to avoid copy
Texture data is often converted before upload, meaning that an array was allocated to perform the conversion into. However, the backend SetData methods were being passed a Span of that data, and the Multithreaded layer does `ToArray()` on it so that it can be stored for later! This method can't extract the original array, so it creates a copy.
This PR changes the type passed for textures to a new ref struct called SpanOrArray, which is backed by either a ReadOnlySpan or an array. The benefit here is that we can have a ToArray method that doesn't copy if it is originally backed by an array.
This will also avoid a copy when running the ASTC decoder.
On NieR this was taking 38% of texture upload time, which it does a _lot_ of when you move between areas, so there should be a 1.6x performance boost when strictly uploading textures. No doubt this will also improve texture streaming performance in UE4 games, and maybe a small reduction with video playback.
From the numbers, it's probably possible to improve the upload rate by a further 1.6x by performing layout conversion on GPU. I'm not sure if we could improve it further than that - multithreading conversion on CPU would probably result in memory bottleneck.
This doesn't extend to buffers, since we don't convert their data on the GPU emulator side.
* Remove implicit cast to array.
2022-10-08 16:04:47 +01:00
|
|
|
|
public void SetData(SpanOrArray<byte> data, int layer, int level, Rectangle<int> region)
|
2022-08-26 03:16:41 +01:00
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureSetDataSliceRegionCommand>().Set(Ref(this), Ref(data.ToArray()), layer, level, region);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
|
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>
2021-08-26 23:31:29 +01:00
|
|
|
|
public void SetStorage(BufferRange buffer)
|
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureSetStorageCommand>().Set(Ref(this), buffer);
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
public void Release()
|
|
|
|
|
{
|
|
|
|
|
_renderer.New<TextureReleaseCommand>().Set(Ref(this));
|
|
|
|
|
_renderer.QueueCommand();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|