Use Quads on OpenGL host when supported. (#2331)

Improves OpenGL performance on FAST RMX and Xenoblade DE/2. Will probably only work on NVIDIA GPUs, but the emulated quads path will still be valid for other GPUs. Note: SLOW RMX gets a bit faster in handheld mode.

I'd recommend checking on platforms without supported host quads to make sure a GL error is actually thrown when attempting GL.Begin(PrimitiveType.Quads)

Ryujinx.Graphics.OpenGL/HwCapabilities.cs

@@ -11,6 +11,7 @@ namespace Ryujinx.Graphics.OpenGL
         private static readonly Lazy<bool> _supportsViewportSwizzle           = new Lazy<bool>(() => HasExtension("GL_NV_viewport_swizzle"));
         private static readonly Lazy<bool> _supportsSeamlessCubemapPerTexture = new Lazy<bool>(() => HasExtension("GL_ARB_seamless_cubemap_per_texture"));
         private static readonly Lazy<bool> _supportsParallelShaderCompile     = new Lazy<bool>(() => HasExtension("GL_ARB_parallel_shader_compile"));
+        private static readonly Lazy<bool> _supportsQuads                     = new Lazy<bool>(SupportsQuadsCheck);
 
         private static readonly Lazy<int> _maximumComputeSharedMemorySize = new Lazy<int>(() => GetLimit(All.MaxComputeSharedMemorySize));
         private static readonly Lazy<int> _storageBufferOffsetAlignment   = new Lazy<int>(() => GetLimit(All.ShaderStorageBufferOffsetAlignment));
@@ -36,6 +37,7 @@ namespace Ryujinx.Graphics.OpenGL
         public static bool SupportsViewportSwizzle           => _supportsViewportSwizzle.Value;
         public static bool SupportsSeamlessCubemapPerTexture => _supportsSeamlessCubemapPerTexture.Value;
         public static bool SupportsParallelShaderCompile     => _supportsParallelShaderCompile.Value;
+        public static bool SupportsQuads                     => _supportsQuads.Value;
         public static bool SupportsNonConstantTextureOffset  => _gpuVendor.Value == GpuVendor.Nvidia;
         public static bool RequiresSyncFlush                 => _gpuVendor.Value == GpuVendor.Amd || _gpuVendor.Value == GpuVendor.IntelWindows || _gpuVendor.Value == GpuVendor.IntelUnix;
         public static bool SupportsMismatchingViewFormat     => _gpuVendor.Value != GpuVendor.Amd && _gpuVendor.Value != GpuVendor.IntelWindows;
@@ -88,5 +90,14 @@ namespace Ryujinx.Graphics.OpenGL
                 return GpuVendor.Unknown;
             }
         }
+
+        private static bool SupportsQuadsCheck()
+        {
+            GL.GetError(); // Clear any existing error.
+            GL.Begin(PrimitiveType.Quads);
+            GL.End();
+
+            return GL.GetError() == ErrorCode.NoError;
+        }
     }
 }

Ryujinx.Graphics.OpenGL/Pipeline.cs

@@ -182,11 +182,11 @@ namespace Ryujinx.Graphics.OpenGL
 
             PreDraw();
 
-            if (_primitiveType == PrimitiveType.Quads)
+            if (_primitiveType == PrimitiveType.Quads && !HwCapabilities.SupportsQuads)
             {
                 DrawQuadsImpl(vertexCount, instanceCount, firstVertex, firstInstance);
             }
-            else if (_primitiveType == PrimitiveType.QuadStrip)
+            else if (_primitiveType == PrimitiveType.QuadStrip && !HwCapabilities.SupportsQuads)
             {
                 DrawQuadStripImpl(vertexCount, instanceCount, firstVertex, firstInstance);
             }
@@ -310,7 +310,7 @@ namespace Ryujinx.Graphics.OpenGL
 
             IntPtr indexBaseOffset = _indexBaseOffset + firstIndex * indexElemSize;
 
-            if (_primitiveType == PrimitiveType.Quads)
+            if (_primitiveType == PrimitiveType.Quads && !HwCapabilities.SupportsQuads)
             {
                 DrawQuadsIndexedImpl(
                     indexCount,
@@ -320,7 +320,7 @@ namespace Ryujinx.Graphics.OpenGL
                     firstVertex,
                     firstInstance);
             }
-            else if (_primitiveType == PrimitiveType.QuadStrip)
+            else if (_primitiveType == PrimitiveType.QuadStrip && !HwCapabilities.SupportsQuads)
             {
                 DrawQuadStripIndexedImpl(
                     indexCount,