* renderer_vulkan: remove out of context SDL_Quit call (fixes#1220)
SDL_Quit must not be called unless the application must exit or SDL is no longer needed
* comment: simplify, forgot init had ref-count
* renderer_vulkan: Disable FIFO when refresh rate is lower than ~60hz
Also disables FIFO when Apple low power mode is enabled, as it can limit the application framerate to 30fps
* renderer_vulkan.cpp: Put `IsLowRefreshRate` into anon namespace + make static
Uses SIMD operations on the RasterizerAccelerated::AnalyzeVertexArray
function, which is hot code. Slightly reduces GPU processing time
on all games.
This idea was suggested by an anonymous contributor.
Co-authored-by: OpenSauce04 <opensauce04@gmail.com>
When the frame limit was set to 0 (unthrottled), the Vulkan present mode would be unintentionally set to FIFO, which caps out at the monitor's refresh rate
* Stereoscopic 3D Enhancements
- Increase maximum 3D depth to 255%
- Fix touch screen to only render 2D when separate window layout is used
- Cleanup some 3D option localizations
* qt: Added warning label below depth slider for values over 100%
* Fixed broken rendering for Interlaced, Reverse Interlaced and Anaglyph options when using 3D with seperate windows
* android: Added warning label below depth slider for values over 100%
* Fixed a bracket and break statement being incorrectly positioned
---------
Co-authored-by: oneup03 <oneup03@gmail.com>
* error checking for layout value from older config
* rename enum and update aspect ratio code
* rewrite LargeFrameLayout to support multiple positions
* add settings for smallscreenposition, fix minsize function
* fixed framebuffer from res scale (screenshots)
* add desktop UI for small screen position
* small screen position submenu on desktop
* fix int-float conversion warning
* rename Above and Below to hopefully fix linux issue
* Add Small Screen Position Setting to android settings menu
* fix sliders to work with floats, mostly
* fix android slider textinput ui
* change None enums in settings and cam_params
* Apply clang-format-18
* SettingsAdapter.kt: Make more null pointer exception resistant
* Updated license headers
* Code formatting nitpicks
* fix bug in main.ui that was hiding menu
* replace default layout with a special call to LargeFrame (like SideBySide does)
* fix bug when "large screen" is actually narrower
* edit documentation for LargeScreenLayout
* update PortraitTopFullFrameLayout to use LargeFrameLayout
* fix unary minus on unsigned int bug
* Applied formatting correction
* Added `const`s where appropriate
* android: Add mention of the bottom-right small screen position being the default
* review fixes + more constants
* refactor all Upright calculations to a reverseLayout method, simplifying code and reducing bugs
* Removed stray extra newline
* SettingsAdapter.kt: Fixed some strange indentation
* Removed unnecessary `if` in favour of direct value usage
---------
Co-authored-by: Reg Tiangha <rtiangha@users.noreply.github.com>
Co-authored-by: OpenSauce04 <opensauce04@gmail.com>
* common/aarch64: Allow generic code generator types
Use the templated `BasicCodeGenerator` type rather than the specialized
`CodeGenerator` type.
Allows `VectorCodeGenerator` to work with these functions.
* common/aarch64: Add `VectorCodeGenerator` to `CallFarFunction`
`VectorCodeGenerator` will always do far-calls since we cannot resolve any absolute addresses here.
* shader_jit_a64: Implement position-independent VectorCodeGenerator
Generates more position-independent assembly to allow for code to be
generated within a resizable vector before copying into executable
memory, allowing for more compact memory allocations and usage rather
than a statically defined worst-case for all-cases.
`VectorCodeGenerator` will need to generate position-independent code
rather than use absolute addresses. Assumes all far function calls in the
case of `VectorCodeGenerator` to use absolute addresses rather than
potentially use a relative `BL` branch after memory relocation.
* Refactor Vulkan stream buffer memory type selection
This is a fix for GPUs with Vulkan V1.1 and V1.2
* add requested refactoring
* clang format
* fix typo
* shader_jit: Add conditional unit-tests
Tests all permutations of X, Y, AND, OR with each possible input value.
* video_core: Fix shader-interpreter conditional-code initialization
Rather than reserving the incoming state of the conditional codes, the
shader-interpreter was setting them both to false. In pretty much all
cases, the initial state of a shaderunit can be zero-initialized
statically. Just running the interpreter shouldn't necessarily reset the
conditional codes though. The JIT loads incoming conditional codes
while the shader-interpreter resets them to false. This makes the
interpreter match the behavior of the shader-jit.
* shader_jit_a64: Fix/optimize conditional evaluation
Fix some of the regressions introduced by the previous optimization.
EOR does not support a constant of `0` in its immediate. In these cases
the COND{0,1} registers can be utilized immediately.
* shader_jit_x64: Fix conditional evaluation extended-bit hazard
The unit test seems to have identified a bug in the x64 jit too. The x64
jit was doing 32-bit comparisons despite the condition flags being 8-bit
values and is sensitive to garbage being in the upper 24 bits of the
register. This is fixed by using the proper 8-bit register types rather
than the 32-bit ones(`eax,`ebx` -> `al`, `bl`).
* shader_jit_x64: Zero-extend conditional-code bytes
`mov` was doing a partial update of bits within the register, allowing
garbage to be introduced in the upper bits of the register.
These conditional tests are a 1:1 translation from the x64 code but do
not have to be. Reference-values are known at emit-time and can be
embedded as an immediate into an `EOR` instruction rather than moved
into a register. The `TST` instruction can be utilized to more optimally
test and update the `EQ`/`NE` status flags.
* Soc and artic_bass: gcc 13+ compatibility fix.
* externals/fmt: update to HEAD fcd3e1e19.
It will fix error.
integer_sequence<bool, (Is == Is)...> [-Werror=tautological-compare]
The updating is helpful and needed.
Fmt has gone through two public versions since its last update
and has fixed many bugs, including new compiler optimizations.
But neither of these two public versions can fix the errors encountered above.
We need to switch to a working version.
It can be fixed after fmt/8e62172.There are still many optimizations,
Such as this one:
Std. h c++23 build fix (# 3856)
And these:
C++23 compatibility: basicstring_view cannot be constructed from nullptr (# 3846)
Fix warning C4702 emitted from format.h (MSVC) (#3866)
Of course, there are other functional improvements as well.
Very helpful.
The selected version is the one that has been checked and works well.
And synchronously updating local code.
* citra_qt/ui: clean up duplicate naming warnings.
* renderer_vulkan: Remove vulkan prefix in SetObjectName
* renderer_vulkan: Rename renderpass cache to render manager
* It is no longer just a cache
* renderer_vulkan: Rewrite descriptor management
* Switch to batched vkUpdateDescriptorSets from cached descriptor sets with templates
* vk_master_semaphore: Remove waitable atomic
* These are buggy on some platforms and regular condition_variables are faster most of the time
* vk_texture_runtime.cpp: remove outdated references
* vk_render_manager: Minor cleanups and rename to RenderManager
* It is no longer just a renderpass cache
* Revert variable name change from render_manager back to renderpass_cache
---------
Co-authored-by: GPUCode <geoster3d@gmail.com>
