Mirrors/yuzu
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

video_core: Move command buffer loop.

Browse source 
This moves the hot loop into video_core. This refactoring shall reduce the CPU overhead of calling ProcessCommandList.
This commit is contained in:
Markus Wick 2018-09-06 15:48:08 +02:00
parent c560043581
commit 0cfb0bacb2
5 changed files with 86 additions and 79 deletions
Download patch file Download diff file Expand all files Collapse all files

4
src/video_core/gpu.h
View file

@ -6,6 +6,7 @@
#include <array>
#include <memory>
#include <vector>
#include "common/common_types.h"
#include "core/hle/service/nvflinger/buffer_queue.h"
#include "video_core/memory_manager.h"
@ -67,6 +68,7 @@ u32 RenderTargetBytesPerPixel(RenderTargetFormat format);
/// Returns the number of bytes per pixel of each depth format.
u32 DepthFormatBytesPerPixel(DepthFormat format);
struct CommandListHeader;
class DebugContext;
/**
@ -115,7 +117,7 @@ public:
~GPU();
/// Processes a command list stored at the specified address in GPU memory.
void ProcessCommandList(GPUVAddr address, u32 size);
void ProcessCommandLists(const std::vector<CommandListHeader>& commands);
/// Returns a reference to the Maxwell3D GPU engine.
Engines::Maxwell3D& Maxwell3D();