From e35ffbbeb0f85f676416fcb8f0bb0207671f379d Mon Sep 17 00:00:00 2001 From: ameerj <52414509+ameerj@users.noreply.github.com> Date: Sun, 30 May 2021 00:53:26 -0400 Subject: glsl: Implement VOTE for subgroup size potentially larger --- src/video_core/renderer_opengl/gl_device.cpp | 1 + src/video_core/renderer_opengl/gl_device.h | 5 +++++ src/video_core/renderer_opengl/gl_shader_cache.cpp | 2 +- 3 files changed, 7 insertions(+), 1 deletion(-) (limited to 'src/video_core') diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp index 071133781..20ea42cff 100644 --- a/src/video_core/renderer_opengl/gl_device.cpp +++ b/src/video_core/renderer_opengl/gl_device.cpp @@ -160,6 +160,7 @@ Device::Device() { has_depth_buffer_float = HasExtension(extensions, "GL_NV_depth_buffer_float"); has_nv_gpu_shader_5 = GLAD_GL_NV_gpu_shader5; has_amd_shader_half_float = GLAD_GL_AMD_gpu_shader_half_float; + warp_size_potentially_larger_than_guest = !is_nvidia && !is_intel; // At the moment of writing this, only Nvidia's driver optimizes BufferSubData on exclusive // uniform buffers as "push constants" diff --git a/src/video_core/renderer_opengl/gl_device.h b/src/video_core/renderer_opengl/gl_device.h index 9b9402c29..ff0ff2b08 100644 --- a/src/video_core/renderer_opengl/gl_device.h +++ b/src/video_core/renderer_opengl/gl_device.h @@ -128,6 +128,10 @@ public: return has_amd_shader_half_float; } + bool IsWarpSizePotentiallyLargerThanGuest() const { + return warp_size_potentially_larger_than_guest; + } + private: static bool TestVariableAoffi(); static bool TestPreciseBug(); @@ -161,6 +165,7 @@ private: bool has_depth_buffer_float{}; bool has_nv_gpu_shader_5{}; bool has_amd_shader_half_float{}; + bool warp_size_potentially_larger_than_guest{}; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 8a052851b..cd11ff653 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -220,7 +220,7 @@ ShaderCache::ShaderCache(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindo .support_gl_nv_gpu_shader_5 = device.HasNvGpuShader5(), .support_gl_amd_gpu_shader_half_float = device.HasAmdShaderHalfFloat(), - .warp_size_potentially_larger_than_guest = true, + .warp_size_potentially_larger_than_guest = device.IsWarpSizePotentiallyLargerThanGuest(), .lower_left_origin_mode = true, .need_declared_frag_colors = true, -- cgit v1.2.3