ryujinx/Ryujinx.Graphics.GAL/Multithreading/ThreadedWindow.cs

35 lines
1.2 KiB
C#
Raw Normal View History

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 22:31:29 +00:00
using Ryujinx.Graphics.GAL.Multithreading.Commands.Window;
using Ryujinx.Graphics.GAL.Multithreading.Model;
using Ryujinx.Graphics.GAL.Multithreading.Resources;
using System;
namespace Ryujinx.Graphics.GAL.Multithreading
{
public class ThreadedWindow : IWindow
{
private ThreadedRenderer _renderer;
private IWindow _impl;
public ThreadedWindow(ThreadedRenderer renderer, IWindow window)
{
_renderer = renderer;
_impl = window;
}
Avalonia UI - Part 1 (#3270) * avalonia part 1 * remove vulkan ui backend * move ui common files to ui common project * get name for oading screen from device * rebase. * review 1 * review 1.1 * review * cleanup * addressed review * use cancellation token * review * review * rebased * cancel library loading when closing window * remove star image, use fonticon instead * delete render control frame buffer when game ends. change position of fav star * addressed @Thog review * ensure the right ui is downloaded in updates * fix crash when showing not supported dialog during controller request * add prefix to artifact names * Auto-format Avalonia project * Fix input * Fix build, simplify app disposal * remove nv stutter thread * addressed review * add missing change * maintain window size if new size is zero length * add game, handheld, docked to local * reverse scale main window * Update de_DE.json * Update de_DE.json * Update de_DE.json * Update italian json * Update it_IT.json * let render timer poll with no wait * remove unused code * more unused code * enabled tiered compilation and trimming * check if window event is not closed before signaling * fix atmospher case * locale fix * locale fix * remove explicit tiered compilation declarations * Remove ) it_IT.json * Remove ) de_DE.json * Update it_IT.json * Update pt_BR locale with latest strings * Remove ')' * add more strings to locale * update locale * remove extra slash * remove extra slash * set firmware version to 0 if key's not found * fix * revert timer changes * lock on object instead * Update it_IT.json * remove unused method * add load screen text to locale * drop swap event * Update de_DE.json * Update de_DE.json * do null check when stopping emulator * Update de_DE.json * Create tr_TR.json * Add tr_TR * Add tr_TR + Turkish * Update it_IT.json * Update Ryujinx.Ava/Input/AvaloniaMappingHelper.cs Co-authored-by: Ac_K <Acoustik666@gmail.com> * Apply suggestions from code review Co-authored-by: Ac_K <Acoustik666@gmail.com> * Apply suggestions from code review Co-authored-by: Ac_K <Acoustik666@gmail.com> * addressed review * Update Ryujinx.Ava/Ui/Backend/OpenGl/OpenGlRenderTarget.cs Co-authored-by: gdkchan <gab.dark.100@gmail.com> * use avalonia's inbuilt renderer on linux * removed whitespace * workaround for queue render crash with vsync off * drop custom backend * format files * fix not closing issue * remove warnings * rebase * update avalonia library * Reposition the Text and Button on About Page * Assign build version * Remove appveyor text Co-authored-by: gdk <gab.dark.100@gmail.com> Co-authored-by: Niwu34 <67392333+Niwu34@users.noreply.github.com> Co-authored-by: Antonio Brugnolo <36473846+AntoSkate@users.noreply.github.com> Co-authored-by: aegiff <99728970+aegiff@users.noreply.github.com> Co-authored-by: Ac_K <Acoustik666@gmail.com> Co-authored-by: MostlyWhat <78652091+MostlyWhat@users.noreply.github.com>
2022-05-15 11:30:15 +00:00
public void Present(ITexture texture, ImageCrop crop, Action<object> swapBuffersCallback)
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 22:31:29 +00:00
{
// If there's already a frame in the pipeline, wait for it to be presented first.
// This is a multithread rate limit - we can't be more than one frame behind the command queue.
_renderer.WaitForFrame();
Avalonia UI - Part 1 (#3270) * avalonia part 1 * remove vulkan ui backend * move ui common files to ui common project * get name for oading screen from device * rebase. * review 1 * review 1.1 * review * cleanup * addressed review * use cancellation token * review * review * rebased * cancel library loading when closing window * remove star image, use fonticon instead * delete render control frame buffer when game ends. change position of fav star * addressed @Thog review * ensure the right ui is downloaded in updates * fix crash when showing not supported dialog during controller request * add prefix to artifact names * Auto-format Avalonia project * Fix input * Fix build, simplify app disposal * remove nv stutter thread * addressed review * add missing change * maintain window size if new size is zero length * add game, handheld, docked to local * reverse scale main window * Update de_DE.json * Update de_DE.json * Update de_DE.json * Update italian json * Update it_IT.json * let render timer poll with no wait * remove unused code * more unused code * enabled tiered compilation and trimming * check if window event is not closed before signaling * fix atmospher case * locale fix * locale fix * remove explicit tiered compilation declarations * Remove ) it_IT.json * Remove ) de_DE.json * Update it_IT.json * Update pt_BR locale with latest strings * Remove ')' * add more strings to locale * update locale * remove extra slash * remove extra slash * set firmware version to 0 if key's not found * fix * revert timer changes * lock on object instead * Update it_IT.json * remove unused method * add load screen text to locale * drop swap event * Update de_DE.json * Update de_DE.json * do null check when stopping emulator * Update de_DE.json * Create tr_TR.json * Add tr_TR * Add tr_TR + Turkish * Update it_IT.json * Update Ryujinx.Ava/Input/AvaloniaMappingHelper.cs Co-authored-by: Ac_K <Acoustik666@gmail.com> * Apply suggestions from code review Co-authored-by: Ac_K <Acoustik666@gmail.com> * Apply suggestions from code review Co-authored-by: Ac_K <Acoustik666@gmail.com> * addressed review * Update Ryujinx.Ava/Ui/Backend/OpenGl/OpenGlRenderTarget.cs Co-authored-by: gdkchan <gab.dark.100@gmail.com> * use avalonia's inbuilt renderer on linux * removed whitespace * workaround for queue render crash with vsync off * drop custom backend * format files * fix not closing issue * remove warnings * rebase * update avalonia library * Reposition the Text and Button on About Page * Assign build version * Remove appveyor text Co-authored-by: gdk <gab.dark.100@gmail.com> Co-authored-by: Niwu34 <67392333+Niwu34@users.noreply.github.com> Co-authored-by: Antonio Brugnolo <36473846+AntoSkate@users.noreply.github.com> Co-authored-by: aegiff <99728970+aegiff@users.noreply.github.com> Co-authored-by: Ac_K <Acoustik666@gmail.com> Co-authored-by: MostlyWhat <78652091+MostlyWhat@users.noreply.github.com>
2022-05-15 11:30:15 +00:00
_renderer.New<WindowPresentCommand>().Set(new TableRef<ThreadedTexture>(_renderer, texture as ThreadedTexture), crop, new TableRef<Action<object>>(_renderer, swapBuffersCallback));
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 22:31:29 +00:00
_renderer.QueueCommand();
}
public void SetSize(int width, int height)
{
_impl.SetSize(width, height);
}
}
}