[video_core/engines/maxwell3d] memory inline DrawState to reduce indirection on hot paths (#3758)

usual indirection removal
helps very slightly to codegen

the idea is basically to reduce the amount of pointer deference overall in the code, and use idiomatic std::variant<>-isms to not rely on vtables/unique_ptr overhead
this should allow the compiler to emit better code
of course it's a tiny optimisation and only CPU side, but allows us to reduce indirection which is almost always a good thing

"but youre passing more parameters to the function!!!" its literally memoized into a register my friend

Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3758
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
This commit is contained in:
lizzie
2026-04-28 17:44:51 +02:00
committed by crueter
parent ed225f8a8b
commit 77decca678
15 changed files with 231 additions and 351 deletions
+6 -7
View File
@@ -13,7 +13,7 @@
#include "core/core.h"
#include "core/core_timing.h"
#include "video_core/dirty_flags.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/gpu.h"
#include "video_core/memory_manager.h"
@@ -26,7 +26,8 @@ namespace Tegra::Engines {
constexpr u32 MacroRegistersStart = 0xE00;
Maxwell3D::Maxwell3D(Core::System& system_, MemoryManager& memory_manager_)
: draw_manager{std::make_unique<DrawManager>(this)}, system{system_}
: draw_manager()
, system{system_}
, memory_manager{memory_manager_}
#ifdef ARCHITECTURE_x86_64
, macro_engine(bool(Settings::values.disable_macro_jit))
@@ -373,8 +374,7 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume
case MAXWELL3D_REG_INDEX(sync_info):
return ProcessSyncPoint();
case MAXWELL3D_REG_INDEX(launch_dma):
return upload_state.ProcessExec(regs.launch_dma.memory_layout.Value() ==
Regs::LaunchDMA::Layout::Pitch);
return upload_state.ProcessExec(regs.launch_dma.memory_layout.Value() == Regs::LaunchDMA::Layout::Pitch);
case MAXWELL3D_REG_INDEX(inline_data):
upload_state.ProcessData(argument, is_last_call);
return;
@@ -386,7 +386,7 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume
case MAXWELL3D_REG_INDEX(tiled_cache_barrier):
return rasterizer->TiledCacheBarrier();
default:
draw_manager->ProcessMethodCall(method, argument);
draw_manager.ProcessMethodCall(*this, method, argument);
break;
}
}
@@ -401,8 +401,7 @@ void Maxwell3D::CallMacroMethod(u32 method, const std::vector<u32>& parameters)
// Execute the current macro.
macro_engine.Execute(*this, macro_positions[entry], parameters);
draw_manager->DrawDeferred();
draw_manager.DrawDeferred(*this);
}
void Maxwell3D::CallMethod(u32 method, u32 method_argument, bool is_last_call) {