Just discovered that stock adreno drivers on a730 had no Immediate VSync mode available, and it was falling back to FIFO, which causes a lot of response delay in Unity games.
But if someone chose Immediate, it prefers low latency above buffering, so Mailbox is his friend, and FIFO is his nemesis.
So i just changed Immediate to fall back to Mailbox first, and only if Mailbox is not available either it falls back to FIFO:
Immediate -> Mailbox -> FIFO;
Mailbox -> FIFO;
FIFO Relaxed -> FIFO.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3746
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
This pr aims to make a first step into giving Eden's Vulkan backend maintenance with better formatting, understanding and reduce the redundancy between some wrong implementations:
-> ProvokingVertex: Has been reworked completely, now the wiring to ExtendedDynamicState3 (ProvokingVertexMode) it works safe, the gating of the extension no longer requires user enabling on UI, it will be enabled automatically based on what features of the extension are available on driver, depending if first or last mode are available and also will be properly set under TrasformFeedback operations; this way we're gonna ensure all drivers, including Android stock drivers on QCOM, Mali and other mobile vendors drivers access correctly to the extension, fixing some graphical issues (flickering textures or wrong sccisors on vertex that required first mode) generated by the missing first mode handling or the proper clearing on pipeline state. This change will increase/ decrease slightly the performance on some games that changes dynamically between first and last modes, but will also ensure a clear path to GPU to use resources smartly, reducing VRAM consumption in PC and MEM/GPU percentage of use on Android (marginal to 5 - 8% approx).
-> Removal VK_EXT_multi_draw: It has passed some months since the first try to implement this feature, but wasn't completed so functionality was null to negative, taking space in source and using small CPU cicles for initialization during device creation that reduced CPU effectiveness by 2 - 5%, aside that, after reviewing the situation of formally introducing handling for multidraws to reduce the floods of constants draw calls into 1 per batch, seems to not pay the overhead when the multidraw capacity is less or equal to 1, aside that, for the time being batching multidraws will introduce indirections on each batch of command draws, that graphical issues will appear and reducing performance can happen, something it was gonna solve. For the time being it's discarded, but, may be analyzed the chance to introduce it properly in a later date.
-> Removal VK_EXT_indexing_descriptor: Currently the implementation of this feature was partial, with not proper handling on layouts; just the initial checks on device creation and chained up to the pNext feature, currently this extension will require a buffer cache and texture cache rework to set layouts and reduce the amount of use of descriptor into making it a less expensive cost effective, aside that to generate a path for the Bindless Texture and Bindless Buffer, allowing to track state of textures in runtime and ensuring the state of some compiled shaders doesn't change if it's no needed, among other benefits, besides that, enabling this feature was only generating innecessary checks on GPU, so consumption would be higher than it should be.
-> Removal of VK_EXT_swapchain_maintenance1: The use of this feature was really conditional to certain support on newer drivers and cards, which wasn't available for all platforms, I concluded that if the support for this instance wasn't really there, the cost effective between the try to initialize it and running it, won't be good as when it didn't exist on our Vulkan device, with also the constant factor of not being completely implemented, right now I'm aiming to reduce the complexity on Vulkan side to keep it as simple as it could for a future video_core rework, which may start somewhere these weeks.
-> Refined the Maintenance features: Meawhile this features are usually inherent to Vulkan core functionality, sometimes drivers doesn't expose them well, which leads to Eden run Vulkan without really noticing the existence of maintenance features on driver to improve stability, so we're tied to declare them and load them when a device creation starts by asking if driver supports and which, currently 1 - 5 are core maintenance features, meanwhile the 7 - 9 are more experimental and not being available everywhere, so right now to help drivers give attention to this features we not only load them, but also calculate with features bits which features inside Maintenance5 are available to use, that was the initial purpose, but we weren't using the full potential of the extension; I made cleaned part of the unneeded log code for the initialization of this Maintenance5 and it's feature flags, but also making an operation to sanitize depth/ stencil when One is not available in swizzle operations, making it easier to draw it; it's also needed to mention that Maintenance9 was removed from source code until we can have better handling on Arrays2D, aside that, the use of the extension was really conditional to certain drivers being capable of use it, so I remove it.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3839
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Co-committed-by: CamilleLaVey <camillelavey99@gmail.com>
Some (Mali) drivers particularly are afraid to unroll loops with more than 7 constant iterations (?); hence manual unrolling is potentially beneficial due to avoiding extra branching + the uniform runtime expectations
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3754
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
PrepareImage() is on a very hot path and previously called SynchronizeAliases() unconditionally. For most images, aliased_images` is empty, so this created unnecessary overhead, now we only synchronize only when image requires it
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3740
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: wildcard <wildcard@eden-emu.dev>
Co-committed-by: wildcard <wildcard@eden-emu.dev>
This Pr is a reply to certain issues found on Android due to the new artificial waits inside Vulkan (Frame Pacing Mode); which caused GPU/CPU desync's even if TimelineSemaphore (Adreno's drivers) does a constant check to retain synchronization with each frame-data, removes the yield() for all platforms (remains the same on PC) and aligns a new way to handle the output of video by using native Android tools, such as AGP, which makes a bridge inside Vulkan to Android's Surface (screen) and reduces not only the latency, but also improves the smoothness of each frame processed; currently we quantize the amount of frame processed by hinting the surface on Android space and adjust the heuristics of the old handling (yuzu) and we link it to screen refresh rate; this way we ensure that even if the game moves below the screen's HZ, we can always pick up the cadence by clamping the duration of each frame and using a chrono function to work as internal fernce if performance goes below the game speed requirment or game's frame rate requirements.
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3735
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: PavelBARABANOV <pavelbarabanov94@gmail.com>
Co-committed-by: PavelBARABANOV <pavelbarabanov94@gmail.com>
#3747 Fixed broken graphics but the game still has a flickering regression on AMD GPU+Windows
This fixes a flickering regression in FFTACTICS. This game on AMD GPU +Windows sensitive to VK_Scheduler barriers. If this game regresses in the future, this is a clue to where to look at first.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3748
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: John <john@eden-emu.dev>
Co-committed-by: John <john@eden-emu.dev>
VK_PIPELINE_STAGE_ALL_COMMANDS_BIT EndRenderPass()
May Fix FFTactics for AMD Windows
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3747
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: John <john@eden-emu.dev>
Co-committed-by: John <john@eden-emu.dev>
Removed the early return in ProcessDirtyRegisters to ensure all dependent dirty flags are set even if the register value hasn't changed.
This fixes flickering and vertex explosions on the Hero's Path in The Legend of Zelda: Breath of the Wild, fixes the teleport bug in The Legend of Zelda: Tears of the Kingdom, fixes the grey Luigi in Luigi's Mansion 3, and may also improve rendering in other games that rely on redundant register writes.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3712
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
- Fixes Ender Magnolia and Demon Slayer freezing on Android.
May be related to this PR or other fixes on the master but:
- Using Cure in Ender Magnolia no longer crashes the game.
- Fixes Demon Slayer vertex explosion on Android.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3728
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: John <john@eden-emu.dev>
Co-committed-by: John <john@eden-emu.dev>
- our logging code was bigger than spdlog itself, why???? just keep it simple
- fix issues when logging before logging system is even started
- removes the "initialized logging twice" issue
- removes uneeded indirection in file logging
- uses direct formatting instead of jumping hoopla-around the fmt::format() ressult
- code duplication and dead code removal as usual
I did explore dup2() but I think it's not worth the hassle
I did try `fwopen()` but it's better if things are just kept as-is.
there is a lot of noise because I removed a bunch of redundant files on logging and just put everything in one file
now normally this wouldn't be a good idea, however consider: the complexity of logging; it's less than 500 lines... does it really need a whole subsystem?!?!?! ITS JUST LOGGING
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3688
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: DraVee <chimera@dravee.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
- OpenGL symbols would still be included in builds without OpenGL, this pr fixes that
- Same goes for Vulkan, but now with `ENABLE_VULKAN`
- Add support to have OpenGL-only builds (why would you do this?)
- Add support for headless runs (yes you could just select NULL backend, but why not compile it headless? :)
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3692
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: DraVee <chimera@dravee.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
regr. Steam Deck
Please, for the love of God, stop saying "YOLO good to merge" after
testers report performance regressions (and promptly get brushed to the
side). Seriously, what the hell?
This reverts commit f8ea09fa0f.
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3704
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: DraVee <chimera@dravee.dev>
This enhances the garbage collection in TextureCache to make it more responsive and reliable during long gameplay sessions.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3652
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
Fix for current crash on master.
Just reverted only the necessary stuff so that PresentManager can hold a reference to khr and resist death upon application hold/restore.
@Lizzie shall judge.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3685
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
MK8D is a big offender, taking up lots of time memcpy'ing and memmov'ing small_vector<> AND to add salt to the wound it doesn't even do heap allocations (no game does I think) - so basically useless waste of compute time in hot path for NO reason :^)
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3641
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: DraVee <chimera@dravee.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Removes some odd #ifdef-ing that just can use a shrimple opaque type.
Also removes nullptr() ctor'ing for vulkan handles and such; it's not incorrect per se like how `void *p = 0;` isn't incorrect, just that, y'know, any static analyzer will go "woah". Also there isn't any guarantee that handles `sizeof(Handle) == sizeof(void*)` so may as well :)
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2971
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Implements push descriptor for compute pipelines along with a bug fix, the increment logic was, offset += sizeof(DescriptorUpdateEntry);
This only advances the byte offset by a single descriptor slot, regardless of the array's size (descriptorCount).Now suppose if a shader utilized an array of descriptors (eg, layout(binding = 0) uniform sampler2D textures[4]) and if this happened to fit within the MaxPushDescriptors limit, the template would consume 4 * sizeof(DescriptorUpdateEntry) bytes, but the offset for the next binding would only advance by 1 slot.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3666
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: wildcard <wildcard@eden-emu.dev>
Co-committed-by: wildcard <wildcard@eden-emu.dev>
Fixes Skyward Sword HD eye gitch and a related MoltenVK crash due to the incorrect output mapping. Verified working on mac and android.
The test in vk_pipeline_cache.cpp is a bit ugly, but it didn't seem worth it to go lambda/macro just to make it look cleaner. Could change if necessary.
Co-authored-by: tarako <none@none.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3637
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: tarako <r76036296@gmail.com>
Co-committed-by: tarako <r76036296@gmail.com>
It makes fine adjustments to the frame pacing, ensuring better stability and precision.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3628
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
This fixes a visual corruption issue that occurred intermittently after loading screens, where some games would start the scene with vertex explosions, artifacts or with all colors blown out, resembling neon.
Among the known games affected by this bug are Mario Kart 8 Deluxe, The Legend of Zelda: Breath of the Wild, The Legend of Zelda: Tears of the Kingdom, Kirby and the Forgotten Land, Luigi's Mansion 3, Xenoblade Chronicles 3 and possibly others as well.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3511
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
in powervr, the enum for cubic filtering is an alias for an ext, however in true powervr fashion they want you to use their stupid VK_IMG
and knowing powevr it probably bugs out if you dont
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3643
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Note to self: you have a 60Hz display, use it
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3633
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Shows a short driver identification string in the Graphics Device combo box (desktop only, Android doesn't need this at all)
Largely meant for debugging, especially macOS. Maybe Windows in the future once MESA begins working on FOSS drivers over there. Linux on ARM platforms too maybe? And Nvidia
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3636
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
This PR aims to return an older way to bind and host vertex/ buffers (via toggle), which had a bunch of indirection and added unnecessary overhead during the drawing phase; current new approach adds just PoD for this operations, which seems to not be acceptable for older turnip drivers.
Meanwhile the performance improvements are gonna be enabled only if the toggle is turned on, it will be required to use newer turnip drivers to make it work (26.0+), default behavior will allow older drivers work as intended.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3621
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: PavelBARABANOV <pavelbarabanov94@gmail.com>
Co-committed-by: PavelBARABANOV <pavelbarabanov94@gmail.com>
(Merge of #3495 + #3108)
This PR works around to simplify math operations on hot pointers inside the access and requests to the cache of buffers and texture cache, removing previous logic of indirection and replaced by a PoD approach.
This will ensure less CPU times spended on the same request and flow directly into another chain of the render, in the same way, command queue currently uses an internal mutex that constraints the flow of data within the GPU threads, we're moving over a single command, I verified to keep using mutexes instead of internal mutex + mutex per operation, which are resolved by themselves.
In simplier words, this aims to improve performance on those games and devices where the waits for next orders on GPU commands were heavier than a single verification.
Co-Authored-by: @CamilleLaVey
Co-Authored-by: @Lizzie
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3579
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
This improves frame pacing control, fixes accumulated drift, and adds the ability to change it while the game is running.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3535
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
This allows users to choose how the emulator manages frame pacing to reduce stuttering and provide a smoother and more consistent frame rate.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3494
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
fixes VVL, but doesnt make it a hard requirement so drivers which cant load it for some reason just, dont
Validation Error: [ VUID-vkCreateInstance-ppEnabledExtensionNames-01388 ] | MessageID = 0xe5e52180
vkCreateInstance(): pCreateInfo->ppEnabledExtensionNames[3] Missing extension required by the instance extension VK_EXT_surface_maintenance1: VK_KHR_get_surface_capabilities2.
The Vulkan spec states: All required extensions for each extension in the VkInstanceCreateInfo::ppEnabledExtensionNames list must also be present in that list (https://docs.vulkan.org/spec/latest/chapters/initialization.html#VUID-vkCreateInstance-ppEnabledExtensionNames-01388)
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3514
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
mainly doing this to reduce memory footprint; we all know how nice ankerl::unordered_dense is
in theory 4x faster - in practice these maps arent that "hot" anyways so not likely to have much perf gained
i just want to reduce mem fragmentation to ease my porting process, plus it helps other platforms as well (ahem weak Mediatek devices) :)
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Co-authored-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3442
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
lizzie