Commit Graph

1514 Commits

Author SHA1 Message Date
Zephyron 9615e910f9 feat(video_core): Implement HybridMemory for advanced Vulkan memory management
Adds a new cross-platform memory management system with enhanced capabilities:
- Fault-managed memory allocation for Linux/Android platforms
- Memory snapshot and differential snapshot support
- Predictive memory reuse tracking for optimized access patterns
- Vulkan compute buffer integration
- User-configurable settings for enabling features

The system integrates with the existing Vulkan renderer to provide more
efficient memory handling, especially for compute-intensive workloads.

Co-authored-by: boss.sfc <boss.sfc@citron-emu.org>
Co-committed-by: boss.sfc <boss.sfc@citron-emu.org>
Signed-off-by: Zephyron <zephyron@citron-emu.org>
2025-04-12 23:41:26 -03:00
Zephyron 0c5659a198 video_core: Enhance Vulkan shader compilation with async threading system
Implement a robust asynchronous shader compilation system inspired by commit
1fd5fefcb17fe7fe65faed1c991fb41db782ab0f. This enhancement provides:

- True multi-threaded shader compilation with atomic status tracking
- Persistent disk caching for faster shader loading
- Command queue system for background processing
- Integration with Citron's scheduler for better resource management
- Parallel shader loading to reduce startup times
- Improved error handling and recovery mechanisms

These changes significantly reduce shader compilation stuttering and improve
overall performance when using asynchronous shaders. The implementation
maintains compatibility with Citron's existing architecture while adding
more robust threading capabilities.

Co-authored-by: boss.sfc <boss.sfc@citron-emu.org>
Co-committed-by: boss.sfc <boss.sfc@citron-emu.org>
Signed-off-by: Zephyron <zephyron@citron-emu.org>
2025-04-12 23:39:24 -03:00
Zephyron b550c6e3cc feat(vulkan): implement enhanced texture and shader management
This commit adds improved Vulkan functionality to the Citron emulator:

- Add thread-safe texture management with automatic error recovery
- Implement shader caching with validation support
- Add robust error handling for Vulkan operations
- Implement platform-specific initialization for Windows, Linux, and Android

These enhancements improve stability when handling texture loading errors
and provide better recovery mechanisms for Vulkan failures.

Co-authored-by: boss.sfc <boss.sfc@citron-emu.org>
Co-committed-by: boss.sfc <boss.sfc@citron-emu.org>
Signed-off-by: Zephyron <zephyron@citron-emu.org>
2025-04-12 23:39:19 -03:00
Zephyron c57a5fef92 WIP: Enhance shader compilation performance and control
This commit adds new settings and optimizations for shader compilation:

- Add new settings:
  - use_enhanced_shader_building: Enable enhanced shader compilation
  - shader_compilation_priority: Control shader compilation priority

- Improve shader compilation performance:
  - Optimize worker thread allocation based on CPU cores
  - Add smarter async shader compilation heuristics
  - Prioritize vertex and fragment shader compilation
  - Add performance tracking and logging

- Add performance monitoring:
  - Track shader compilation times
  - Log slow shader compilations
  - Monitor async shader compilation statistics

This is a work in progress commit. Further optimizations and refinements
will be needed based on testing and feedback.

Signed-off-by: Zephyron <zephyron@citron-emu.org>
2025-04-12 23:37:11 -03:00
MrPurple666 ce1a1912ae RenderVulkan Experiments: DO-NOT-MERGE: Only recreate when necessary and NEAREST 2025-04-12 22:54:32 -03:00
MrPurple666 2a793d610d RenderVulkan Experiments: DO-NOT-MERGE: Introduce functionality to frame interpolation
The functionality acts much more like a real-time downscale since the FSR1 used by Yuzu is not compatible with a real interpolation and is just upscale and downscale.
2025-04-12 22:54:32 -03:00
Pavel Barabanov 38737f26e5 Optimize SpiriV output 2025-04-12 22:54:32 -03:00
Pavel Barabanov be2e838f0d FRAME_SKIPPING and FRAME_INTERPOLATION only android and delete "enable" 2025-04-12 22:54:32 -03:00
Pavel Barabanov d25bea5762 Android: add FRAME_SKIPPING and FRAME_INTERPOLATION 2025-04-12 22:54:32 -03:00
Zephyron 45683a4034 video_core/vulkan: Improve texture format conversion handling
Refactors and improves the texture format conversion system in the Vulkan
renderer:

- Adds proper sRGB to linear conversion for depth formats
- Improves shader accuracy for ABGR8 SRGB to D24S8 conversion
- Adds gamma correction and proper depth range clamping
- Moves GetSupportedFormat implementation to header
- Cleans up format conversion switch statement
- Removes redundant format conversion paths

The changes improve accuracy when converting between color and depth
formats, particularly for sRGB sources. The shader improvements ensure
proper gamma correction and depth range handling.

Technical changes:
- Improves sRGB to linear conversion in fragment shader
- Adds proper depth value clamping
- Consolidates format conversion logic
- Removes duplicate GetSupportedFormat implementation
2025-04-11 17:05:41 -03:00
Zephyron 3e835ac3aa video_core: Add new shader format conversion pipelines
Adds several new shader-based format conversion pipelines to support additional
texture formats and operations:

- RGBA8 to BGRA8 conversion
- YUV420/RGB conversions
- BC7 to RGBA8 decompression
- ASTC HDR to RGBA16F decompression
- RGBA16F to RGBA8 conversion
- Temporal dithering
- Dynamic resolution scaling

Updates the texture cache runtime to handle these new conversion paths and adds
helper functions to check format compatibility for dithering and scaling
operations.

The changes include:
- New shader files and CMake entries
- Additional conversion pipeline setup in BlitImageHelper
- Extended format conversion logic in TextureCacheRuntime
- New format compatibility check helpers
2025-04-11 17:05:30 -03:00
Zephyron 5cb3153f15 video_core: Add sRGB to D24S8 depth-stencil conversion support
Implements conversion from sRGB color formats to D24S8 depth-stencil format
in the Vulkan renderer. This change includes:

- New fragment shader convert_abgr8_srgb_to_d24s8.frag that handles proper
  sRGB to linear conversion before depth calculation
- Added shader to CMake build system
- Extended BlitImageHelper with new conversion pipeline and methods
- Updated texture cache to handle sRGB to D24S8 format conversion paths

The conversion properly handles sRGB color space by first converting to
linear space before calculating luminance values for the depth component,
while preserving alpha channel data for the stencil component.
2025-04-11 17:05:20 -03:00
crueter 9cef9f5dce Update fmt to allow for v11
Signed-off-by: swurl <swurl@swurl.xyz>
2025-04-10 01:11:22 +02:00
JPikachu 52eab40fc8 Implement LogicOP Workaround for Vulkan on AMD GPUs
This fixes black sewers in Paper Mario: TTYD
Credit: Antique - [Sudachi] Dev (https://sudachi.emuplace.app/)
2025-04-08 19:17:28 +02:00
darktux 492d3856e8 Vulkan validation error fix.
Different image usage flags between image creation and image view
creation.
2025-04-04 03:40:48 +02:00
darktux 2032d035a3 Removed telemetry and anonymized SCM (git) strings 2025-04-04 03:40:47 +02:00
liamwhite 5f2369bba6 general: workarounds for SMMU syncing issues (#12749) 2024-02-27 15:42:15 +01:00
liamwhite a918ce6fe9 vk_rasterizer: flip scissor y on lower left origin mode (#13122) 2024-02-27 15:40:33 +01:00
Liam 9db29b4e20 settings: remove global override for smash on amdvlk 2024-02-26 11:16:18 -05:00
Liam e15b1021a4 video_core: make gpu context aware of rendering program 2024-02-26 11:16:14 -05:00
Narr the Reg 091abdc5cb Merge pull request #13075 from liamwhite/mali-having-a-bad-time
shader_recompiler: throw on missing geometry streams in geometry shaders
2024-02-22 11:30:26 -06:00
FearlessTobi 4d55e58ee4 scope_exit: Make constexpr
Allows the use of the macro in constexpr-contexts.
Also avoids some potential problems when nesting braces inside it.
2024-02-19 16:00:46 +01:00
Liam 749f0564ae shader_recompiler: throw on missing geometry streams in geometry shaders 2024-02-19 00:34:00 -05:00
Liam ba17a8c2b8 gpu: dependency-inject scaling/antialiasing filter state for capture layers 2024-02-09 09:20:53 -05:00
Liam ce72818075 nvnflinger/gpu: implement applet capture 2024-02-09 09:20:53 -05:00
Liam 1c32e4157c nvnflinger/gpu: implement blending 2024-02-09 09:20:53 -05:00
Liam 85abd0243e video_core: defensively program around unmapped device pointers 2024-02-09 09:20:53 -05:00
Fernando Sahmkow 80d505b5cf Buffer Cache: Refactor to use Range sets instead 2024-02-05 11:06:52 +01:00
Fernando Sahmkow 4d97c54676 VideoCore: Move Slot Vector to Common 2024-02-04 20:01:47 +01:00
liamwhite 3868b44440 Merge pull request #12885 from Moonlacer/eclipse-fix
structured_control_flow: Add Samsung Proprietary Driver ID to Reorder Pass
2024-02-02 09:08:54 -05:00
Moonlacer 309ca2bd3b Clang Fix 2024-02-01 18:15:21 -06:00
Moonlacer 31200d0c7a Add Samsung Proprietary Driver ID to Reorder Pass
For RDNA-based Samsung Xclipse GPUs
2024-02-01 17:53:26 -06:00
Liam 5ec76db392 nvnflinger/gpu: implement layer stack composition 2024-01-31 11:27:21 -05:00
Liam f727f13cb9 renderer_vulkan: implement layer stack composition 2024-01-31 11:27:21 -05:00
Liam 2a314f9e2e renderer_opengl: split up blit screen resources into antialias and window adapt passes 2024-01-31 11:27:21 -05:00
Liam cc8493ea2e renderer_vulkan: convert FSR to graphics pipeline 2024-01-31 11:27:21 -05:00
Liam bd9c228844 renderer_opengl: split out FXAA 2024-01-31 11:27:20 -05:00
Liam 86bfa837fc renderer_vulkan: split up blit screen resources into separate antialias and window adapt passes 2024-01-31 11:27:20 -05:00
Liam f50296ab23 renderer_vulkan: isolate FXAA from blit screen 2024-01-31 11:27:20 -05:00
Liam cd8049a5e2 video_core: consistently account for resolution scaling when rendering 2024-01-31 11:27:20 -05:00
Liam 84678c7014 video_core: simplify accelerated surface fetch and crop handling between APIs 2024-01-31 11:27:20 -05:00
liamwhite bc857b4b5a Merge pull request #12439 from FireBurn/vkresult
Simplify VkResult lookup
2024-01-29 09:08:32 -05:00
liamwhite 0ab550c726 Merge pull request #12499 from Kelebek1/time
Rework time services
2024-01-25 14:19:01 -05:00
Kelebek1 52c35521bd Rework time service to fix time passing offline. 2024-01-24 04:26:55 +00:00
liamwhite 5a6b3a0125 Merge pull request #12579 from FernandoS27/smmu
Core: Implement Device Mapping & GPU SMMU
2024-01-22 10:55:39 -05:00
Mike Lothian 0a3c09f2f6 Simplify VkResult lookup 2024-01-22 03:10:43 +00:00
Fernando Sahmkow 241c64e43b Core: Eliminate core/memory dependancies. 2024-01-18 21:12:30 -05:00
Fernando Sahmkow b206089ea7 Core: Clang format and other small issues. 2024-01-18 21:12:30 -05:00
Fernando Sahmkow f11218a03e SMMU: Implement physical memory mirroring 2024-01-18 21:12:30 -05:00
Fernando Sahmkow 9db159da71 SMMU: Initial adaptation to video_core. 2024-01-18 21:12:30 -05:00