Henrik RydgårdandClaude Opus 5.5 f35ddce87f Vulkan: Do the GE-exact FIXA/FIXB blend math in highp ...

The fragment shader's default precision is lowp, which NVIDIA's Vulkan driver
runs as fp16, and the product (2c+1)(2f+1) reaches 511*511, past fp16's range.
Split/Second's bloom saturated whole channels; its two bloom dumps go from MSE
2354/9305 to 80/153, the same as OpenGL.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

2026-10-09 18:43:26 +02:00

..

DepalettizeShaderCommon.cpp

Remove the concept of scaled/offset depth buffers.

2026-06-10 10:41:30 +02:00

DepalettizeShaderCommon.h

…

DepthBufferCommon.cpp

Depth readback: Scale by 65536, and decode in full precision

2026-10-08 10:36:19 +02:00

DepthRaster.cpp

DepthRaster: fix host buffer overflows in the depth raster queue

2026-08-30 22:33:01 +02:00

DepthRaster.h

DepthRaster: fix host buffer overflows in the depth raster queue

2026-08-30 22:33:01 +02:00

DepthSwizzle.h

Core: Let 16-bit CPU accesses through the depth mirrors see the GE's layout

2026-10-08 18:37:23 +02:00

Draw2D.cpp

Vulkan: Look up multisampled framebuffers through a CLUT by their first sample

2026-10-09 15:49:15 +02:00

Draw2D.h

Vulkan: Look up multisampled framebuffers through a CLUT by their first sample

2026-10-09 15:49:15 +02:00

DrawEngineCommon.cpp

Splines: Skip the second vertex decode of tessellated curves

2026-10-08 18:12:17 +02:00

DrawEngineCommon.h

Splines: Skip the second vertex decode of tessellated curves

2026-10-08 18:12:17 +02:00

FragmentShaderGenerator.cpp

Vulkan: Do the GE-exact FIXA/FIXB blend math in highp

2026-10-09 18:43:26 +02:00

FragmentShaderGenerator.h

Remove some D3D9 leftovers

2025-06-10 16:59:14 +02:00

FramebufferManagerCommon.cpp

Merge pull request #22472 from hrydgard/msaa-fixes

2026-10-09 16:53:35 +02:00

FramebufferManagerCommon.h

Vulkan: Look up multisampled framebuffers through a CLUT by their first sample

2026-10-09 15:49:15 +02:00

GPUDebugInterface.cpp

Centralize a disassembly utility function between the debuggers

2026-08-11 09:08:52 +02:00

GPUDebugInterface.h

Remove the concept of scaled/offset depth buffers.

2026-06-10 10:41:30 +02:00

GPUStateUtils.cpp

Keep 5551 stencil marks when an EQUAL 0 test fails

2026-10-09 12:05:20 +02:00

GPUStateUtils.h

Compute FIXA/FIXB blends like the GE

2026-10-09 09:09:36 +02:00

ImageCommon.h

Add a common TextureAlpha enum, replacing CheckAlphaResult and ReplacedTextureAlpha.

2026-06-09 11:25:50 +02:00

IndexGenerator.cpp

GPU: Fix draw engine buffer overruns and stale vertex data

2026-09-29 10:33:32 -06:00

IndexGenerator.h

Fast culling: Use the worldviewproj matrix to avoid the need for planes

2026-05-23 14:40:15 +02:00

PostShader.cpp

Bound two values that texture packs and shader inis supply

2026-09-02 18:06:47 +02:00

PostShader.h

Rework constant buffer loading, barrier fix

2026-04-23 13:32:06 +02:00

PresentationCommon.cpp

GPU: Draw frames displayed from RAM in non-buffered mode

2026-09-29 13:17:01 -06:00

PresentationCommon.h

GPU: Draw frames displayed from RAM in non-buffered mode

2026-09-29 13:17:01 -06:00

ReinterpretFramebuffer.cpp

Do some cleanup of #includes in GPU

2024-12-18 13:57:26 +01:00

ReinterpretFramebuffer.h

Some include cleanup, delete some obsolete code

2024-07-25 14:52:41 +02:00

ReplacedTexture.cpp

Allow specifying a cap on the number of threads used in ParallelLoop

2026-10-09 14:33:58 +02:00

ReplacedTexture.h

TextureReplacer: Fix stale and shared lookup results

2026-09-29 13:28:21 -06:00

ShaderCommon.cpp

…

ShaderCommon.h

GPU: Delete copy operations on classes that own resources

2026-09-29 10:33:32 -06:00

ShaderId.cpp

Don't write fragment depth when the depth test is effectively off

2026-10-08 16:16:24 +02:00

ShaderId.h

Don't write fragment depth when the depth test is effectively off

2026-10-08 16:16:24 +02:00

ShaderUniforms.cpp

GPU: Saturate the lighting pow's exponent below 512, as the GE does

2026-10-05 13:32:24 -06:00

ShaderUniforms.h

GPU: Remove leftovers from hardware tessellation

2026-09-30 11:00:02 -06:00

SoftwareLighting.h

…

SoftwareTransformCommon.cpp

Drop 3D rectangles with a vertex out of range, like the software renderer

2026-10-09 09:09:36 +02:00

SoftwareTransformCommon.h

Software transform clipping: split clipped quads like the GE, cull beyond far first

2026-10-08 16:16:24 +02:00

SplineCommon.cpp

SoftGPU: Skin the x, y and z components together in vector lanes

2026-10-09 09:07:12 +02:00

SplineCommon.h

Splines: Vectorize the GE-exact spline evaluation

2026-10-08 19:33:20 +02:00

StencilCommon.cpp

Rebuild the stencil from alpha after reinterpreting a framebuffer

2026-10-08 10:45:57 +02:00

StencilCommon.h

…

TextureCacheCommon.cpp

Vulkan: Look up multisampled framebuffers through a CLUT by their first sample

2026-10-09 15:49:15 +02:00

TextureCacheCommon.h

GPU: Delete copy operations on classes that own resources

2026-09-29 10:33:32 -06:00

TextureDecoder.cpp

VSH prep: Free up kernel memory by not storing the PPGe atlas in it

2026-08-19 18:45:52 +02:00

TextureDecoder.h

Rename the get-memory-pointer functions to make it clear where CPU exceptions can happen.

2026-08-12 14:06:16 +02:00

TextureReplacer.cpp

TextureReplacer: Fix stale and shared lookup results

2026-09-29 13:28:21 -06:00

TextureReplacer.h

GPU: Delete copy operations on classes that own resources

2026-09-29 10:33:32 -06:00

TextureScalerCommon.cpp

Allow specifying a cap on the number of threads used in ParallelLoop

2026-10-09 14:33:58 +02:00

TextureScalerCommon.h

[GPU] Make static and const methods if possible

2024-04-05 17:04:31 +03:00

TextureShaderCommon.cpp

GPU: Fix assorted leaks, null derefs and small rendering bugs

2026-09-29 10:33:32 -06:00

TextureShaderCommon.h

Fix initialization bug in TextureCacheCommon

2026-06-14 14:02:09 +02:00

TransformCommon.cpp

GPU: Saturate the lighting pow's exponent below 512, as the GE does

2026-10-05 13:32:24 -06:00

TransformCommon.h

GPU: Saturate the lighting pow's exponent below 512, as the GE does

2026-10-05 13:32:24 -06:00

VertexDecoderArm64.cpp

Vertex decoder: JIT morph on arm64

2026-09-21 16:51:00 -06:00

VertexDecoderArm.cpp

Remove obsolete vertex decoder code and options

2026-07-14 17:26:00 +02:00

VertexDecoderCommon.cpp

GPU: Fix draw engine buffer overruns and stale vertex data

2026-09-29 10:33:32 -06:00

VertexDecoderCommon.h

GPU: Fix draw engine buffer overruns and stale vertex data

2026-09-29 10:33:32 -06:00

VertexDecoderHandwritten.cpp

Vertex decoder: fix the handwritten decoders, and test them too

2026-09-21 11:29:40 -06:00

VertexDecoderHandwritten.h

Add another handwritten vertex decoder

2024-05-11 10:00:39 +02:00

VertexDecoderLoongArch64.cpp

Vertex decoder: bring the riscv64 and loongarch64 JITs in line

2026-09-21 13:48:56 -06:00

VertexDecoderRiscV.cpp

Vertex decoder: bring the riscv64 and loongarch64 JITs in line

2026-09-21 13:48:56 -06:00

VertexDecoderX86.cpp

Vertex decoder: make the JIT match the step functions

2026-09-21 11:29:40 -06:00

VertexReader.h

Disable anisotropic filtering for all flat draws.

2026-06-08 15:28:58 +02:00

VertexShaderGenerator.cpp

Vulkan: Interpolate texture coordinates and colors at the centroid

2026-10-09 15:49:15 +02:00

VertexShaderGenerator.h

Remove some D3D9 leftovers

2025-06-10 16:59:14 +02:00