summaryrefslogtreecommitdiff
path: root/src/shader_recompiler (follow)
Commit message (Collapse)AuthorAgeFilesLines
...
| * shader_recompiler: remove barriers in conditional control flow when device ↵Gravatar Liam2023-06-105-0/+51
| | | | | | | | lacks support
* | shader_recompiler: translate f64 to f32 when unsupported on hostGravatar Liam2023-06-105-0/+191
|/
* video_core: Enable support_descriptor_aliasing on Turnip, disable storage ↵Gravatar bunnei2023-06-031-0/+10
| | | | atomic otherwise.
* Avoid using VectorExtractDynamic for subgroup mask on Adreno GPUsGravatar Billy Laws2023-06-032-1/+18
| | | | This crashes their shader compiler for some reason.
* Implement scaled vertex buffer format emulationGravatar Billy Laws2023-06-035-49/+75
| | | | These formats are unsupported by mobile GPUs so they need to be emulated in shaders instead.
* shader_recompiler: fix copy-paste errorGravatar Liam2023-05-261-1/+1
|
* Fix Tears of the Kingdom flickering clouds and depths.Gravatar Kelebek12023-05-112-12/+3
|
* Define SampleMask as an arrayGravatar Kelebek12023-04-302-2/+4
|
* shader_recompiler: Use vector arithmetic rather than component-wise in ↵Gravatar Wollnashorn2023-04-081-18/+9
| | | | | | ImageGatherSubpixelOffset Should be more efficient and better readable
* video_core: Enable ImageGather with subpixel offset on IntelGravatar Wollnashorn2023-04-081-3/+3
|
* shader_recompiler: Add subpixel offset for correct rounding at `ImageGather`Gravatar Wollnashorn2023-04-083-0/+72
| | | | | | | On AMD a subpixel offset of 1/512 of the texel size is applied to the texture coordinates at a ImageGather call to ensure the rounding at the texel centers is done the same way as in Maxwell or other Nvidia architectures. See https://www.reedbeta.com/blog/texture-gathers-and-coordinate-precision/ for more details why this might be necessary. This should fix shadow artifacts at object edges in Zelda: Breath of the Wild (#9957, #6956).
* general: fix spelling mistakesGravatar Liam2023-03-122-41/+41
|
* buffer_cache: Add logic for non-NVN storage buffer trackingGravatar ameerj2023-02-251-1/+7
|
* Merge pull request #9588 from liamwhite/bylaws-revertsGravatar liamwhite2023-02-194-12/+6
|\ | | | | Revert "shader_recompiler: Align SSBO offsets to meet host requirements"
| * Revert "shader_recompiler: Align SSBO offsets to meet host requirements"Gravatar Liam2023-01-074-12/+6
| | | | | | | | This reverts commit 8804a4eb23e0c4f3e4bab03dee7c204bd38bf21e.
* | glsl_emit_context: Remove redeclarations of gl_SampleID and gl_SampleMaskGravatar ameerj2023-02-081-6/+0
| | | | | | | | These built-ins seem to be available without needing to be declared for fragment shaders, similar i.e. to gl_FragDepth
* | shader_recompiler/value.h: Remove lingering references to S32Gravatar ameerj2023-02-041-11/+0
| |
* | Merge pull request #9703 from ameerj/txq-msGravatar liamwhite2023-02-015-18/+51
|\ \ | | | | | | shaders: Fix TXQ with MSAA textures
| * | spirv: Fix TXQ with MSAA texturesGravatar ameerj2023-01-293-8/+19
| | |
| * | emit_glasm_image: Fix TXQ with MSAA texturesGravatar ameerj2023-01-291-1/+9
| | |
| * | emit_glsl_image: Implement TXQ with MSAA texturesGravatar ameerj2023-01-291-9/+23
| | | | | | | | | | | | Also fixes for texture buffers, which do not have mips eithers.
* | | Move to Clang Format 15Gravatar Levi Behunin2023-01-293-10/+10
|/ / | | | | | | | | | | Depends on https://github.com/yuzu-emu/build-environments/pull/69 clang-15 primary run
* | Merge pull request #9699 from ameerj/texture-pass-descGravatar liamwhite2023-01-291-2/+9
|\ \ | | | | | | texture_pass: Fix texture descriptors comparisons
| * | texture_pass: Fix texture descriptors comparisonsGravatar ameerj2023-01-281-2/+9
| | |
* | | Merge pull request #9698 from ameerj/texture-pass-handleGravatar liamwhite2023-01-291-7/+7
|\ \ \ | | | | | | | | texture_pass: Refactor texture handle retrieval
| * | | texture_pass: Refactor texture handle retrievalGravatar ameerj2023-01-281-7/+7
| | | |
* | | | Merge pull request #9694 from ameerj/txq-mipsGravatar liamwhite2023-01-2911-29/+37
|\ \ \ \ | |/ / / |/| | | shader_recompiler: TXQ: Skip QueryLevels when possible
| * | | shader_recompiler: TXQ: Skip QueryLevels when possibleGravatar ameerj2023-01-2811-29/+37
| |/ /
* | | Merge pull request #9687 from ameerj/ogl-shader-msGravatar bunnei2023-01-284-33/+46
|\ \ \ | | | | | | | | glasm, glsl: Implement multisampled Image Fetch
| * | | emit_glsl_image: Fix ImageFetch for MSAA texturesGravatar ameerj2023-01-281-6/+11
| | | |
| * | | glasm: Add MS sampler typesGravatar ameerj2023-01-272-5/+8
| | | |
| * | | glsl: Add MS sampler typesGravatar ameerj2023-01-271-22/+27
| |/ /
* | | Merge pull request #9682 from ameerj/shader-s32Gravatar bunnei2023-01-2813-46/+19
|\ \ \ | |/ / |/| | shader_recompiler: Remove S32 IR type
| * | shader_recompiler: Remove S32 IR typeGravatar ameerj2023-01-2513-46/+19
| | | | | | | | | | | | | | | | | | The frontend IR opcodes do not distinguish between signed and unsigned integer types. Fixes broken shaders when IR validation/graphics debugging is enabled for shaders that used BitCastS32F32
* | | spirv: fix multisampled image fetchGravatar Liam2023-01-234-2/+16
|/ /
* / Avoid OOB array access reading passthrough attr maskGravatar Billy Laws2023-01-071-1/+1
|/ | | YFC 1.5 extended the size of the varying mask used to hold passthrough attrs without considering this
* Run clang-formatGravatar Billy Laws2023-01-055-23/+33
|
* shader_recompiler: Fix shuffle partitioning for >64 invoc-per-subgroup GPUsGravatar Billy Laws2023-01-051-30/+28
| | | | The existing implementation only supports 64 invoc-per-subgroup GPUs, and misbehaves on adreno when invocations need to be split into 4 emulated subgroups.
* shader_recompiler: Add support for lowering geometry passthroughGravatar Billy Laws2023-01-052-40/+67
| | | | Reuses most of the existing code for generating the gl_Layer passthrough. Fixes geometry in Nier: Automata on GPUs without HW passthrough support.
* shader_recompiler: Align SSBO offsets to meet host requirementsGravatar Billy Laws2023-01-054-6/+11
| | | | We can take advantage of SSBO addresses being passed in a constant bufer to account for the extra alignment requirements in the shader itself.
* shader_recompiler: SPIRV: Only enable int64 feature when supportedGravatar Billy Laws2023-01-051-1/+1
|
* shader_recompiler: Add comparison operators to descriptor typesGravatar Billy Laws2023-01-051-0/+12
|
* Vulkan: Add a workaround for input_position on Adreno driversGravatar Billy Laws2023-01-054-11/+41
| | | | Adreno drivers will crash compiling geometry shaders if the input position is not wrapped in a gl_in struct.
* Video_core: Address feedbackGravatar Fernando Sahmkow2023-01-049-0/+39
|
* ShaderCompiler: Inline driver specific constants.Gravatar Fernando Sahmkow2023-01-032-1/+34
|
* MacroHLE: Final cleanup and fixes.Gravatar Fernando Sahmkow2023-01-011-2/+2
|
* MacroHLE: Add OpenGL SupportGravatar Fernando Sahmkow2023-01-012-1/+13
|
* MacroHLE: Add HLE replacement for base vertex and base instance.Gravatar Fernando Sahmkow2023-01-0112-6/+91
|
* Merge pull request #7450 from FernandoS27/ndc-vulkanGravatar liamwhite2022-12-173-3/+5
|\ | | | | Vulkan: Add support for VK_EXT_depth_clip_control.
| * Vulkan: Add support for VK_EXT_depth_clip_control.Gravatar FernandoS272022-12-133-3/+5
| |
HIC SVNT DRACONES