[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>

2 files changed, 12 insertions, 12 deletions

diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 26826a8f78..70f244809f 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -245,7 +245,7 @@ void RasterizerOpenGL::PrepareDraw(bool is_indexed, Func&& draw_func) {
 
     SyncState();
 
-    const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
+    const auto& draw_state = maxwell3d->draw_manager.draw_state;
 
     const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(draw_state.topology);
     BeginTransformFeedback(pipeline, primitive_mode);
@@ -260,12 +260,12 @@ void RasterizerOpenGL::PrepareDraw(bool is_indexed, Func&& draw_func) {
 
 void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
     PrepareDraw(is_indexed, [this, is_indexed, instance_count](GLenum primitive_mode) {
-        const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
-        const GLuint base_instance = static_cast<GLuint>(draw_state.base_instance);
-        const GLsizei num_instances = static_cast<GLsizei>(instance_count);
+        const auto& draw_state = maxwell3d->draw_manager.draw_state;
+        const GLuint base_instance = GLuint(draw_state.base_instance);
+        const GLsizei num_instances = GLsizei(instance_count);
         if (is_indexed) {
-            const GLint base_vertex = static_cast<GLint>(draw_state.base_index);
-            const GLsizei num_vertices = static_cast<GLsizei>(draw_state.index_buffer.count);
+            const GLint base_vertex = GLint(draw_state.base_index);
+            const GLsizei num_vertices = GLsizei(draw_state.index_buffer.count);
             const GLvoid* const offset = buffer_cache_runtime.IndexOffset();
             const GLenum format = MaxwellToGL::IndexFormat(draw_state.index_buffer.format);
             if (num_instances == 1 && base_instance == 0 && base_vertex == 0) {
@@ -302,7 +302,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
 }
 
 void RasterizerOpenGL::DrawIndirect() {
-    const auto& params = maxwell3d->draw_manager->GetIndirectParams();
+    const auto& params = maxwell3d->draw_manager.indirect_state;
     buffer_cache.SetDrawIndirect(&params);
     PrepareDraw(params.is_indexed, [this, &params](GLenum primitive_mode) {
         if (params.is_byte_count) {
@@ -358,12 +358,12 @@ void RasterizerOpenGL::DrawTexture() {
 
     SyncState();
 
-    const auto& draw_texture_state = maxwell3d->draw_manager->GetDrawTextureState();
+    const auto& draw_texture_state = maxwell3d->draw_manager.draw_texture_state;
     const auto& sampler = texture_cache.GetGraphicsSampler(draw_texture_state.src_sampler);
     const auto& texture = texture_cache.GetImageView(draw_texture_state.src_texture);
 
     const auto Scale = [&](auto dim) -> s32 {
-        return Settings::values.resolution_info.ScaleUp(static_cast<s32>(dim));
+        return Settings::values.resolution_info.ScaleUp(s32(dim));
     };
 
     Region2D dst_region = {
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index cf62b02c97..9275dbf01c 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -25,7 +25,7 @@
 #include "shader_recompiler/frontend/maxwell/control_flow.h"
 #include "shader_recompiler/frontend/maxwell/translate_program.h"
 #include "shader_recompiler/profile.h"
-#include "video_core/engines/draw_manager.h"
+#include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/kepler_compute.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/memory_manager.h"
@@ -357,7 +357,7 @@ GraphicsPipeline* ShaderCache::CurrentGraphicsPipeline() {
     const auto& regs{maxwell3d->regs};
     graphics_key.raw = 0;
     graphics_key.early_z.Assign(regs.mandated_early_z != 0 ? 1 : 0);
-    graphics_key.gs_input_topology.Assign(maxwell3d->draw_manager->GetDrawState().topology);
+    graphics_key.gs_input_topology.Assign(maxwell3d->draw_manager.draw_state.topology);
     graphics_key.tessellation_primitive.Assign(regs.tessellation.params.domain_type.Value());
     graphics_key.tessellation_spacing.Assign(regs.tessellation.params.spacing.Value());
     graphics_key.tessellation_clockwise.Assign(
@@ -397,7 +397,7 @@ GraphicsPipeline* ShaderCache::BuiltPipeline(GraphicsPipeline* pipeline) const n
     // If games are using a small index count, we can assume these are full screen quads.
     // Usually these shaders are only used once for building textures so we can assume they
     // can't be built async
-    const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
+    const auto& draw_state = maxwell3d->draw_manager.draw_state;
     if (draw_state.index_buffer.count <= 6 || draw_state.vertex_buffer.count <= 6) {
         return pipeline;
     }