Reduces the page entries from 32 bytes to 8 and rewrites `VirtualBuffer` to be more efficient in memory usage and specifically for large zero regions.
The page table will now only reserve 1GiB instead of 4GiB and of this memory it should only use at most ~8MiB.
This PR has the side effect of using Eden on Windows on low memory systems much more plausible since it would previously require ~10GiB of committable memory at front (despite using ~5-6 at most, inadvertently stalling other processes) where as now it should only require around the amount that it'll actually use.
Co-authored-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4219
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Storage buffers can span GPU pages that aren't contiguous in host memory, but the buffer cache assumed one contiguous buffer per binding, so anything crossing a mapping boundary read the wrong bytes. Multi-range binding resolves the real segments and presents them to the shader as one buffer, aliasing their memory into a sparse VkBuffer or gathering them into a copy, all of this was mostly fixed by making a path to use sparse on buffers (the same way as texture cache has it) and retrieve properly the mapping ranges to the virtual pages, including the actual structure on the use of binding sparse.
Meanwhile this fixes Monster Hunter Sunbreak z-fighting, bad performance and mostly vertex explosions (coming from the contiguous buffer read from the shader's game) it's only the basis for a further investigation towards how are we gonna treat phi calculations, optimized them and mostly eradicate the currently excess of exposure on the textures where the lightning trace should not being reflected.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4362
Reviewed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.
-------------------
Problem: In cyrilic idioms, Absolum won't past beyond 1st area (intro miniboss).
From the user panic i`ve extracted these strings:
> GlobalState.request_next()
> save_state_to_saveGame(...)
> platform.ProcessSave()
> bin_serialize_to_file(...)
> StandaloneTypeModel.Write(SaveGame, ...)
> Write(SaveGame.AdventureSaveSlot, ...)
> Write(PlayerState, ...)
> ProtoBuf.ProtoWriter.WriteString(...) + 0xcc
> ProtoBuf.Helpers.DebugAssert(...)
> Debug.Fail(...)
By debugging ProtoBuf.ProtoWriter.WriteString(...) argument it was found that the string Отлично! (Great! expression for combos between 6 and 30 iirc) was reaching a size comparison of different values and causing an assert in guest side..
It seems the game tries to save highest combo level. I've managed to skip this bug by doing a perfect combo for entire level, so that the problematic word never was generated or attempted to be saved.
After further research, the issue chain was identified: Game maths literal string size via two ways and compare, and one of them was getting replaced by a wrong value due to Eden letting the literal string be interpreted as a real instruction.
Logic added:
const bool pair = decltype(l)::ExtractValue(raw) & 1;
const bool fixed_rt2 = decltype(rt2)::ExtractValue(raw) == 0b11111;
return pair || fixed_rt2;
- pair checks bit 21 (L & 1): register-pair instructions actually use Rt2, so they remain accepted.
- For single-register forms, require Rt2 == 31. The table word fails this check, so NCE leaves it intact.
Sources
https://www.scs.stanford.edu/~zyedidia/arm64/stxrb.htmlhttps://www.scs.stanford.edu/~zyedidia/arm64/stxp.htmlhttps://github.com/qemu/qemu/blob/master/target/arm/tcg/a64.decode#L344
Altough the solution is a case specific guard, it is data agnostic: This is a true fix to avoid a specific case in which a literal may be interpreted as a valid instruction, without interfering with other true instruction cases.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4369
Reviewed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Replaces all instances of ankerl's unordered map/set with boost's
`unordered_flat_*` classes. This uses std::hash since boost::hash is
actually a lot slower.
Also adds an abstraction layer in `Common` so future changes are quicker
and easier.
Other implementation details:
- ankerl provided hash specializations for tuple and pair, so those were
ported here
- std::erase_if doesn't work on boost, so just used the ADL'd erase_if
This should be about equal or superior performance as unordered_dense for everything except iteration.
Signed-off-by: crueter <crueter@eden-emu.dev>
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.
-------------------
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4326
Reviewed-by: Lizzie and Samuel <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Signed-off-by: lizzie <lizzie@eden-emu.dev>
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.
-------------------
logs normally are like this:
```
08-27 02:58:48.188 12511 4743 W YuzuNative: [18534.062505] Shader <Warning> shader_recompiler/frontend/maxwell/translate/impl/move_special_register.cpp:141:Read: (STUBBED) SR_WSCALEFACTOR_XY
08-27 02:58:48.188 12511 4743 W YuzuNative: [18534.062513] Shader <Warning> shader_recompiler/frontend/maxwell/translate/impl/move_special_register.cpp:144:Read: (STUBBED) SR_WSCALEFACTOR_Z
08-27 02:58:48.188 12511 4743 W YuzuNative: [18534.062520] Shader <Warning> shader_recompiler/frontend/maxwell/translate/impl/vote.cpp:50:VOTE_vtg: (STUBBED) called
08-27 02:58:48.188 12511 4743 W YuzuNative: [18534.062523] Shader <Warning> shader_recompiler/frontend/ir/ir_emitter.cpp:267:GetFlowTest: (STUBBED) FCSM_TR
```
bunch of redundant info imo
instead they should just be
```
08-27 02:58:48.188 12511 4743 W YuzuNative: Shader shader_recompiler/frontend/maxwell/translate/impl/move_special_register.cpp:141:Read: (STUBBED) SR_WSCALEFACTOR_XY
08-27 02:58:48.188 12511 4743 W YuzuNative: Shader shader_recompiler/frontend/maxwell/translate/impl/move_special_register.cpp:144:Read: (STUBBED) SR_WSCALEFACTOR_Z
08-27 02:58:48.188 12511 4743 W YuzuNative: Shader shader_recompiler/frontend/maxwell/translate/impl/vote.cpp:50:VOTE_vtg: (STUBBED) called
08-27 02:58:48.188 12511 4743 W YuzuNative: Shader shader_recompiler/frontend/ir/ir_emitter.cpp:267:GetFlowTest: (STUBBED) FCSM_TR
```
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4310
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
PR/Commit Owner: CamilleLaVey
This PR prevents the recompilation of shaders when a new session has been started, based on data found on the recompiler, each new session does still recompile already cached shader cache, reducing the reaction and speed of games by at certain level, this prevent the warm-up with post sessions, only needed the initial warm-up and works globally for all games.
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4294
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
this PR reworks IPS parser to be less stupid
what do i mean by this? well generally give it a bit of
love so it doesn't do a lot of unsound allocations
also simplify the logic greatly (and use memcmp() string idiom
instead of trusting the compiler so blindly...)
no this doesn't mean to uber optimize IPS to handle 999 gb/s
it's more so it doesn't outright crash with edge cases
as the previous codebase was quite spaghetty
also the major overhead is obviously the vector shenanigans
and the I/O -- but thats out of scope
Test that IPSwitch mods still properly work WITH ANY GAME
IF THERE IS ANY REGRESSION IN SOME GAME/MOD THEN
TELL ME
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3911
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.
-------------------
This one deserves a long story, but imma try to resume:
During investigating Absolum 1.2 black screen of death upon loading intro video, i've discovered guest was aborting for failing to allocate room for the video.
By logging everything prior to crash and decoding guest side instructions managed to confirm its media allocator was reading data faster than it was updating free available bucket list.
Since the IStorage::Read was happening 247 times before the crash, i've decided to add a very small sleep there, and boom, not only Absolum but some other titles got the same issue fixed.
But i was unsatisfied with the sleep and kept tracking guest instructions upstream in order to find a sync point for the read worker and the memory allocation update. But unfortunately the media allocator helpers live in guest, accessing memory directly via MMU, so any sync signaling would need to come from some dynarmic hack.
It's been 6 days now, so i've decided to polish the sleep: Moved it upstream to where i could have access for proper predicate, and added machinery to service and nvhost to support that. Now the sleep is restricted only for nvdec istorage reads. Any other reads will flow normally.
TL;DR: currently our code is so blazing async that guest is capable to request reads before its very self refresh it have freed room to do so. The sleep accepted as broadly stable was 600 us (MICROseconds), and it affects ONLY nvdec chunk reading.
Reports confirm that now videos are smoother now.
Code was polished at my knowledge limits.
Mostly machinery to track when a request comes from a process with nvdec active, and is istorage read.
I can provide more details if it comes to be needed.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4316
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: Samuel <lizzie@eden-emu.dev>
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.
-------------------
- Dynamic shader cache reloading capability for qlaunch
- Multi-process improvements (thanks to @frank1734), instead of just using the main application we now respect caller process (also did it for HID devices while I was at it)
- Layer stack masks & shared buffer screenshot for video core, added different masks (screenshot, recording, etc.) this was discovered as an issue due to how in qlaunch the transition was not right between applications. May not be perfect but also fixes screenshots while using qlaunch
- Reworked overlay display management (input and visibility) - instead of random numbers as I've previously did, I decided to add an AppletZIndex enum for better readability. Also split capability of input by touch and gamepad.
- etc.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4238
Reviewed-by: Samuel <lizzie@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
lizzie