paraLLEl N64 RDP – Android support and Intel iGPU improvements – What you should know (and what to expect)

June 2, 2020August 17, 2021 Libretro Team

Ridge Racer 64 running on Parallel RDP on an Android phone (with RetroArch)

Themaister wrote an article a few days ago talking in-depth about all the work that has gone into ParaLLEl RDP since launch.

Two of the important things discussed in this article were:
* Intel iGPU performance
* Android support

What you might not have realized from reading the article is that with the right tweaks, you can already get ParaLLEl RDP to run reasonably well. As indicated in the article he wrote, Themaister will be looking at WSI Vulkan issues specifically related to RetroArch since there definitely do seem to be some issues that have to be resolved. In the meantime, we have to resort to some workarounds. Workarounds or not, they will do the job for now.

How to install and set it up

In RetroArch, go to Online Updater.
(If you have paraLLEl N64 already installed) – Select ‘Update Installed Cores’. This will update all the cores that you already installed.
(If you don’t have paraLLEl N64 installed already) – go to ‘Core Updater’, and select ‘Nintendo – Nintendo 64 (paraLLEl N64)’.
Now start up a game with this core.
Go to the Quick Menu and go to ‘Options’. Scroll down the list until you reach ‘GFX Plugin’. Set this to ‘parallel’. Set ‘RSP plugin’ to ‘parallel’ as well.
For the changes to take effect, we now need to restart the core. You can either close the game or quit RetroArch and start the game up again.

Intel iGPU

What you should do for optimum performance right now:

For Intel iGPU, I have found that what makes the biggest difference by far (on Windows 10 at least) is to run it in windowed mode instead of fullscreen. Fullscreen mode will have horribly crippled performance by comparison.

Performance

Once you have done this, the performance will actually not be that far behind with a run-off-the-mill iGPU from say a 2080 Ti (in asynchronous mode). Sure, it’s still a bit slower by about ~30fps, but it’s no longer the massive gulf in performance it was before where even Angrylion was beating ParaLLEl RDP in the performance department.

With synchronous, the difference between say a 2080 Ti and an iGPU should be a bit more pronounced.

Hopefully in future RetroArch versions, it will no longer be necessary to have to resort to windowed mode for good performance with Intel iGPUs. For now, this workaround will do.

Android

What you should do for optimum performance right now:

Turn vsync off. Go to Settings -> Video -> Synchronization, and make sure that ‘Vertical Sync (Vsync)’ is disabled.

NOTE: It is imperative that you turn V-Sync off for now. If not, performance will be so badly crippled that even Angrylion will be faster by comparison. Fortunately, there will be no noticeable screen tearing even with Vsync disabled right now.

Performance

I tested ParaLLEl RDP on two devices:

Nvidia Shield TV (2015)
Samsung Galaxy S10 Plus (2019) [European Exynos model]

NOTE: The European model of the Galaxy S10 Plus used here has the Samsung Exynos SoC (System-On-A-Chip). Generally these perform worse than the US models of the Galaxy phones, which use a Qualcomm Snapdragon SoC instead. You should therefore expect significantly better performance on a US model.

Performance on Shield TV

Here are some rough performance figures for the Nvidia Shield TV –

Title	Performance
Mortal Kombat Trilogy	87 to 94fps
Yoshi’s Story	99fps
Doom 64	90 to 117fps
Tetris 64	117fps
Starcraft 64	177fps

It’s hard to put an exact number on other games, but just from a solely gameplay-focused perspective, you can get a near-locked framerate with games like Legend of Zelda: Ocarina of Time and Super Mario 64 if you run the PAL versions (which limit the framerate to 50fps instead of 60fps with NTSC versions). There might still be the odd frame drop in certain graphics intensive scenes but nothing too serious.

Similarly, games like 1080 Snowboarding drop below fullspeed with the NTSC version, but running them with the PAL version is nearly a locked framerate in all but the most intensive scenes.

Performance on Samsung Galaxy S10 Plus

Performance on a high-end 2019 phone like the Galaxy S10 Plus can tend to be more variable, probably because of the aggressive dynamic throttling being done on phones. Sometimes performance would be a significant step above the Shield TV where it could run NTSC versions of games like Legend of Zelda: Ocarina of Time and Super Mario 64 at fullspeed with no problem (save for the very odd frame drop here and there in very rare scenes), and then at other times it would perform similarly to a Shield TV. Your mileage may vary there.

Conclusions

Overall, it’s clear that certain battles have to be won on the Vulkan side, especially when it comes down to having to disable vsync at all so far for acceptable performance.

We’d like to learn more from people who have a Samsung Galaxy S20 or a similar high end phone released in 2020. Even a Snapdragon version of the S10 Plus would produce better results than what we see here.

So, Low-Level N64 emulation, is it attainable on Android? Yes, with the proper Vulkan extensions, and provided you have a reasonably modern and fast high end phone. The Shield TV is also a decent mid-range performer considering its age. Far from every game runs at fullspeed yet but the potential is certainly there for this to be a real alternative to HLE based N64 emulation on Android as hardware grows more powerful over the years.

FAQ

Some specific issues should be addressed –

Game compatibility is significantly lower on Android right now

The mupen64plus-core part of ParaLLEl N64 is older than the one found in Mupen64plus next. While on PC this is not so much of an issue because of the generally mature (but slower) Hacktarux dynarec, on ARM platforms it is a different story since new_dynarec was in a premature state back then. Not only that, LLE RDP + RSP plugin compatibility with new_dynarec was not even a consideration back then. So some games might not work at all right now with Parallel RDP+RSP on Android.

ParaLLEl N64 will likely receive a mupen64plus-core update soon, and Mupen64Plus Next might also in the near future get ParaLLEl RDP + ParaLLEl RSP support. So this situation will sort itself out.

You get a display error showing ‘ERR’ on your Android device

The Vulkan driver for your GPU is likely missing these two Vulkan extensions, which ParaLLEl RDP requires.

VK_KHR_8bit_storage
VK_KHR_16bit_storage

(Intel iGPU) Performance is halved (or more) in fullscreen mode

Known issue, read above. These issues have been identified and it’s a matter of finding the appropriate solution for these issues.

paraLLEl-RDP update

May 31, 2020August 17, 2021 Maister

Since the paraLLEl-RDP rewrite was unleashed upon the world, a fair bit of work has gone into it. Mostly performance related and working around various drivers.

Rendering bug fixes

Unsurprisingly, some bugs were found, but very few compared to what I expected. All the rendering bugs were fortunately rather trivial in nature, and didn’t take much effort to debug. I can only count 3 actual bugs. To be a genuine bug, the issue must be isolated to paraLLEl-RDP. Core bugs are unfortunately quite common and a lot of core bugs were mistaken as RDP ones.

Mega Man 64 – LODFrac in Cycle 1

The RDP combiner can take the LOD fractional value as inputs to the combiner. However, the initial implementation only considered that Cycle 0 would observe a valid LODFrac value. This game however, uses LODFrac in Cycle 1, and that case was completely ignored. Fixing the bug was as simple as consider that case as well, and the RDP dump validated bit-exact against Angrylion. I believe this also fixed some weird glitching in Star Wars – Naboo. At least it too passed bit-exact after this fix was in place.

Mario Tennis crashes – LoadTile overflow

Some games, Mario Tennis in particular will occasionally attempt to upload textures with broken coordinates. This is supposed to overflow in a clean way, but I missed this case, and triggered an “infinite” loop with 4 billion texels being updated. Needless to say, this triggered GPU crashes as I would exhaust VRAM while spamming an “infinite” loop with memory allocations. Fairly simple fix once I reproduced it. I believe I saw these crashes in a few other games as well, and it’s probably the same issue. Haven’t seen any issues since the fix.

Perfect Dark logo transition

Not really an RDP rendering issue, but VI shenanigans. This was a good old case of a workaround for another game causing issues. When the VI is fed garbage input, we should render black, but that causes insane flickering in San Francisco Rush, since it strobes invalid state every frame. Not entirely sure what’s going on here (not impossible it’s a core bug …), but I applied another workaround on top of the workaround. I don’t like this 🙁 At least the default path in the VI implementation is to do the expected thing of rendering black here, and parallel-n64 opts into using weird workarounds for invalid VI state.

Core bugs

Right now, the old parallel-n64 Mupen core is kind of the weakest link, and almost all issues people report as RDP bugs are just core bugs. I’ll need to integrate this in a newer Mupen core and see how that works out.

Improving compatibility with more Vulkan drivers

As mentioned in my last post, a workaround for lack of VK_EXT_external_memory_host was needed, and I implemented a fairly complex scheme to deal with this in a way that is not horribly slow. Effectively, we now need to shuffle memory back and forth between two views of RDRAM, the CPU-owned RDRAM, and GPU-owned RDRAM. The implementation is quite accurate, and tracks writes on a per-byte basis.

The main unit of work submitted to the GPU is a “render pass” (similar in concept to a Vulkan render pass). This is a chunk of primitives which all render to the same addresses in RDRAM and which do not have any feedback effect, where texture data is sampled from the frame buffer region being rendered to. A render pass will have a bunch of reads from RDRAM at the start of the render pass, where frame buffer data is read, along with all relevant updates to TMEM. All chunks of RDRAM which might be read, will be copied over to GPU RDRAM before rendering. We also have a bunch of potential writes after the render pass. These writes must eventually make their way back to CPU RDRAM. Until we drain the GPU for work completely, any write made by the RDP “wins” over any writes made by the CPU. During the “read” phase of the render pass, we can selectively copy bytes based on the pending writemask we maintain on the GPU. If there are no pending writes by GPU, we optimize to a straight copy.

As for performance, I get around 10-15% FPS hit on NVIDIA with this workaround. Noticeable, but not crippling.

Android

Android SoCs do not always support cache-coherency with the GPU, so that’s added complexity. We have to carefully flush caches and invalidate caches on CPU side after we write to GPU RDRAM and before we read from it respectively. I also fixed a bunch of issues with cache management in paraLLEl-RDP which would never happen on a desktop system, since everything is essentially cache coherent.

With these fixes, paraLLEl-RDP runs correctly on at least Galaxy S9/S10 with Android 10 and Mali GPUs, and the Tegra in Shield TV. However, the support for 8/16-bit storage is still very sparse on Android, and I couldn’t find a single Snapdragon/Adreno GPU supporting it, oh well. One day Android will catch up. Don’t expect any magic for the time being w.r.t. performance, there are some horrible performance issues left which are Android specific outside the control of paraLLEl-RDP, and need to be investigated separately.

Fixing various performance issues

The major bulk of the work was fixing some performance issues which would come up in some situations.

Building a profiler

To drill down into these issues, I needed better tooling to be able to correlate CPU and GPU activity. This was a good excuse to add such support into Granite, which is paraLLEl-RDP’s rendering backend, Beetle HW Vulkan’s backend, and the foundation of my personal Vulkan rendering engine. Google Chrome actually has a built-in profile UI frontend in chrome://tracing which is excellent for ad-hoc use cases such as this. Just dump out some simple JSON and off you go.

To make a simple CPU <-> GPU profiler all you need is Vulkan timestamp queries and VK_EXT_calibrated_timestamps to improve accuracy of CPU <-> GPU timestamp correlation. I made use of the “pid” feature of the trace format to show the different frame contexts overlapping each other in execution.

Anyone can make these traces now by setting environment variables: PARALLEL_RDP_BENCH=1 GRANITE_TIMESTAMP_TRACE=mytrace.json, then load the JSON in chrome://tracing.

Why is Intel Mesa much slower than Intel Windows?

This was one of the major questions I had, and I figured out why using this new tool. In async mode, performance just wouldn’t improve over sync mode at all. The reason for this is that swap buffers in RetroArch would completely stall the GPU before completing (“refresh” in the trace). I filed a Mesa bug for this. I’ll need to find a workaround for this in RetroArch. With a hacky local workaround, iGPU finally gives a significant uplift over just using the CPU in this case. Trace captured on my UHD 620 ultrabook which shows buggy driver behavior. Stalling 6 ms in the main emulation thread is not fun. 🙁

Fixing full GPU stalls, or, why isn’t Async mode improving performance?

This was actually a parallel-n64 bug again. To manage CPU <-> GPU overlap, the Vulkan backend uses multiple frame contexts, where one frame on screen should correspond with one frame context. The RDP integration was notified too often that a frame was starting, and thus would wait for GPU work to complete far too early. This would essentially turn Async mode into Sync mode in many cases. Overall, fixing this gained ~10-15% FPS on my desktop systems.

Be smarter about how we batch up work for the GPU – fixing stutters in Mario Tennis

Mario Tennis is pretty crazy in how it renders some of its effects. The hazy effect is implemented with ~50 (!) render passes back to back each just rendering one primitive. This was a pathological case in the implementation that ran horribly.

The original design of paraLLEl-RDP was for larger render passes to be batched up with a sweet spot of around 1k primitives in one go, and each render pass would correspond to one vkQueueSubmit. This assumption fell flat in this case. To fix this I rewrote the entire submission logic to try to make more balanced submits to the GPU. Not too large, and not too small. Tiny render passes back-to-back will now be batched together into one command buffer, and large render passes will be split up. The goal is to submit a meaningful chunk of work to the GPU as early as possible, and not hoard tons of work while the GPU twiddles its thumbs. This is critically important for Sync mode I found, because once we hit a final SyncFull opcode, we will need to wait for the GPU to complete all pending work. If we have already submitted most of the relevant work, we won’t have to wait as long. Overall, this completely removed the performance issue in Mario Tennis for me, and overall performance improved by a fair bit. > 400 VI/s isn’t uncommon in various games now on my main system. RDP overhead in sync mode usually accounts for 0.1 ms – 0.2 ms per frame or something like that, quite insignificant.

Performance work left?

I think paraLLEl-RDP itself is in a very solid place performance-wise now, the main issues are drilling down various WSI issues that plague Intel iGPU and Android, which I believe is where we lose most of the performance now. That work would have to go into RetroArch itself, as that’s where we handle such things.

Overall, remember that accurate LLE rendering is extremely taxing compared to HLE rendering pixel-for-pixel. The amount of work that needs to happen for a single pixel is ridiculous when bit-exactness is the goal. However, shaving away stupid, unnecessary overhead has a lot of potential for performance uplift.

RetroArch 1.8.8 released!

May 27, 2020August 17, 2021 Libretro Team

RetroArch 1.8.8 has just been released.

Grab it here.

Read our latest Libretro Cores Progress Report blog post here. It’s an exhaustive list, and especially the older consoles have received a lot of new cores and improvements.

Remember that this project exists for the benefit of our users, and that we wouldn’t keep doing this were it not for spreading the love with our users. This project exists because of your support and belief in us to keep going doing great things. If you’d like to show your support, consider donating to us. Check here in order to learn more. In addition to being able to support us on Patreon, there is now also the option to sponsor us on Github Sponsors! You can also help us out by buying some of our merch on our Teespring store!

Highlights

Add option to sort playlists after name truncation (Ozone)

A couple of users have complained about a feature we made in 1.8.7 (Fix sidebar playlist sort order when ‘Truncate Playlist Names’ is enabled). This new addition makes the new sidebar playlist sorting behaviour optional via a Sort Playlists After Name Truncation setting under User Interface > Appearance. When disabled, playlists will be sorted the old way (according to file name), not by display name.

Localization – big updates and crowdsourced

A new language has been added, Slovakian. And plenty of the existing languages have received big updates as far as localization goes.

But by far the biggest change is our transition to Crowdin. This allows non-programmers to more easily contribute localization changes/additions to RetroArch. You can see the completion status of the various languages on our Crowdin page.

Improved shader preset dirs

1.8.8 restores the original behavior of the “Save Shader Preset As” option, and improves the cycling of shaders by allowing to cycle the shaders on fallback directories if the Video Shader directory does not contain any preset.

Input Remapping Fixes

In 1.8.7, pressing RetroPad Start to reset a core input remap to the default setting did not work correctly – analog stick inputs get the wrong defaults, and inputs that are left undefined by the core are not set to the proper RARCH_UNMAPPED value.

1.8.8 fixes the issue.

RetroArch WiiU: Gamepad hotplugging support, theoretical multi-gamepad support

1.8.8 adds support for hotplugging WiiU Gamepads. Critically for users, this makes the driver no longer assume a Gamepad is present, so when it’s broken or out of battery or missing; the first Wiimote gets slot 1 instead (helpful when All Users Control Menu is off). I [quarkawesome] also made it check Gamepad channel 2 – while it’s impossible to connect a second Gamepad on a retail console, the code to do it still appears to be there. If that feature ever becomes a thing through CFWs, it’ll work here.

RetroArch P2 – New SDK/toolchain – big improvements

fjtrujy spent a lot of time adapting RetroArch PlayStation2 to the latest PS2 SDK. RetroArch PS2 is now being built with a modern version of the GCC compiler, and certain cores are already seeing massive speedups as a result.

I honestly didn't expect this, but it looks that new releases also have better performance in some cores. #QuickNes is one of them #RetroArch ⁦@libretro⁩ pic.twitter.com/P8iCbXvvuh

— Francisco J. Trujillo (@fjtrujy) May 21, 2020

As can be seen by the tweet, QuickNES went from 255fps with the old SDK to 429fps with the new SDK. This makes the core more than fast enough to use runahead – on a PlayStation2 of all things!

A newer C/C++ toolchain also will make it much easier to port over software to PS2, it was previously quite difficult to port C++ cores over to PS2.

fjtrujy also added Theodore to the list of cores supported.

Changelog

What you’ve read above is just a small sampling of what 1.8.8 has to offer. There might be things that we forgot to list in the changelog listed below, but here it is for your perusal regardless.

1.8.8

AUDIO/JACK: Fix regression introduced after 1.8.4 – would hang at startup
CHEEVOS: Disable hardcore when cheats are enabled
CHD: Return false when special track cannot be found
DISCORD/MATCHMAKING: Fix Discord ‘Ask To Join’ functionality
FILE PATH: Various file path handling optimisations
FONT: Fix Arabic, Chinese and Korean font rendering
INPUT MAPPING/REMAPPING: Restore broken ‘reset to default’ functionality with RetroPad ‘start’ button
INPUT MAPPING/REMAPPING: Fix ‘reset to default’ action for analog sticks and undefined core inputs
LIBRETRO: Add new message extension allowing for richer messages
LOCALIZATION: Update Arabic translation
LOCALIZATION: Update Chinese (Simplified) translation
LOCALIZATION: Update Chinese (Traditional) translation
LOCALIZATION: Update German translation
LOCALIZATION: Update Greek translation
LOCALIZATION: Update Spanish translation
LOCALIZATION: Update French translation
LOCALIZATION: Update Italian translation
LOCALIZATION: Update Japanese translation
LOCALIZATION: Update Korean translation
LOCALIZATION: Update Dutch translation
LOCALIZATION: Update Polish translation
LOCALIZATION: Update Portuguese Brazilian translation
LOCALIZATION: Update Russian translation
LOCALIZATION: Update Turkish translation
LOCALIZATION: Update Vietnamese translation
LOCALIZATION: Add Slovak translation
MENU: Small buffer optimizations
MENU/THUMBNAILS/BUGFIX: Fix heap-use-after-free error
MENU/OZONE: Add option to sort playlists after name truncation
MENU/OZONE/ANDROIDTV: Default to Ozone menu driver
MENU/OZONE/ANDROID: Gamepad-like devices default to Ozone now (Shield Portable)
NETPLAY: Lower announcement rate
OVERLAYS: Fix memory leak when loading overlays
SHADER PRESETS: Improved shader preset dirs
TIME/DATE: Enable configuration of date seperator in clock and runtime ‘last played’ displays
VITA: Fix upside-down vertical games
UWP: Enable playlist and savefile compression by default (because of slow file I/O)
VIDEO/WIDGETS: Fix overlapping text when simultaneous pop-up notifications and core/shader messages are being displayed
WIIU: Gamepad hotplugging support
WIIU: Theoretical multi-gamepad support
X11: Fix crash in x11_display_server_get_screen_orientation
X11/XSHM: Allow X11/XHSM video driver to operate without SHM extension
X11/XSHM: Fix compatibility with X11 input driver
XVIDEO: Fix keyboard input initialization
XVIDEO/XWAYLAND: Fix XVideo support on xwayland (by supporting I420 and YZ12)

Libretro Cores Progress Report – May 27, 2020

May 27, 2020August 17, 2021 Libretro Team

Yesterday marked RetroArch’s 10 Year Anniversary date. Today, we’re not only releasing a new RetroArch version, but we’re letting you know all the changes that have been made to the libretro cores since the last progress report.

Our last core progress report was on April 2, 2020. Below we detail the most significant changes to all the Libretro cores we and/or upstream partners maintain. We are listing changes that have happened since then.

How to update your cores in RetroArch

There are two ways to update your cores:

a – If you have already installed the core before, you can go to Online Updater and select ‘Update Installed Cores’.

b – If you haven’t installed the core yet, go to Online Updater, ‘Core Updater’, and select the core from the list that you want to install.

Final Burn Neo

Description: Multi-system arcade emulator

Add Neo Geo Pocket Color support
Latest updates from upstream

Flycast

Description: Sega Dreamcast/NAOMI emulator

Naomi/AW widescreen integrated cheats (KNIGHTS OF VALOUR THE 7 SPIRITS, Metal Slug 6, Toy Fighter, Dolphin Blue)
DSP: Proper MIXS input shift. Fixes Grandia 2 missing sound effects
DSP: fix output shift. Fix wrungp ear rape (NAOMI game)
Fix for: [Bug]Super Street Fighter II X for Matching Service (Japan) – Disable BIOS region patching
Fix for: Samurai Spirits – Frame-skipping issues before reset or changing the option – don’t reset frameskip to 0 at init
Haiku: Fix build
Rewrite nvidia jetson nano build
Make threaded rendering the default on all platforms. Synchronous mode enabled unless LOW_END is defined
Info and warning for xBRZ upscaling core option
ARM64: check CpuRunning at end of each timeslice. fix hang when exiting service menu in kofnw (NAOMI game)
NAOMI: wrungp inputs. support inverted axis for NAOMI
Libretro: Improve context request
Libretro: fix input descriptor L2/R2 mixup
NAOMI: Add Tokyo Bus Guide support
Log VMU files loading
CUSTOM TEXTURES: support JPEG format. Get rid of libpng and use stb_image log error if naomi eeprom save fails
Fix mouse state not being updated
PVR/NAOMI: update palette when PAL_RAM_CTRL is updated – fixes wrong palette in Gun Survivor 2 score screen
REND: Use original palette data to compute palette hash
PVR: textured background plane – fixes Who Wants To Be a Millionaire

DOSbox Core

Description: DOSbox core

Latest updates from upstream
Add option for using 2axis joystick even when only one port is connected – this fixes input problems in “Super Off Road”, but other games might also need this.
Add option for controlling log verbosity level – Since we can log to stdout now, it makes sense. Also, some frontends might not have configurable log verbosity levels.
Add option for printing log output to stdout/stderr –
Useful if the frontend’s logging output is unreliable or too noisy (or
both) and we only want to see log output from the core. RetroArch does
have a configuration option for controlling frontend and core log output
seperately, but it’s bugged.
Enable Voodoo on all platforms and regardless of fakesdl – Software-based Voodoo emulation doesn’t require SDL anymore so it should build fine everywhere.
Don’t claim there’s Voodoo2 emulation – Oops. 12MB doesn’t actually mean Voodoo2. The current code only emulates the original Voodoo. The 12MB setting is just a non-standard memory configuration for the Voodoo 1.
Add support for changing current core option values programmatically.
This is a hack. The libretro API does not actually support this. We
achieve this by replacing all current values with a bogus one to force
the frontend to forget the currently selected value, since it doesn’t
match the bogus one. We then submit the correct values again, but with
the default value set to the value we want to switch to. This forces
the frontend to switch to that value because the bogus value is now
gone and thus not a valid value anymore. Finally, we submit the values
again but with the initial default value (we only want to change the
current value, not set a new default.) The frontend will not switch the
current value, as the values themselves have not changed, just the
default has.

RetroArch seems to be well-behaved here and does the correct thing.
Other frontends might not play ball though.
Hook up 3dfx core options
Vita: Fix dynarec, fix build
Add build options to make bassmidi and fluidsynth optional
Fix ARM dynarec
Correct cdrom sector size field length according to docs.
Refactor input mapper –
Code should be simpler to understand now. This also fixes a bug where
inputs on the second port weren’t working before. Mouse emulation is now
possible on both ports and the default emulated mouse buttons have been
swapped with the speed modifier buttons (L/R are now mouse buttons,
L2/R2 the speed modifiers.) This is a saner default since not all
controllers have L2/R2 buttons.
GHA: Support macOS 10.13 by building with GCC instead of XCode Clang
Fix floppy image file size detection oopsie
Improve disk control related code and move it to its own source file
Make image file extension comparisons case-insensitive –
This fixes the issue where loading an image that has an upper-case
extension (like “.CUE” instead of “.cue”) results in dosbox mounting the
image itself without going through the libretro disk control interface.

We add new case conversion funtions for this (in the new util.h/util.cpp
files) because the existing conversion functions provided either by
dosbox or libretro-common are crap and we’re smarter than everybody
else.
Add libretro disk control interface disk labels support –
Only retro_get_image_label_t for now. Leave retro_set_initial_image_t
and retro_get_image_path_t undefined as it’s not clear what the use
would be in case of DOS games, especially since we don’t handle m3u
files yet.
Fix disk_replace_image_index always reporting failure –
This silly mistake caused the “Failed to append disc” error message in
RetroArch.
Add core option for setting the free space when auto-mounting drive C
Ensure overlay mount path ends with a dir separator –
Otherwise dosbox will write data in bogus directories in the overlay.
Add option for mounting the executable’s parent dir as drive C –
Some pre-installed games expect to be installed in C:\GAMEDIR rather
than directly in C:\. This option allows these games to run without
having to modify their configuration files first.
Enable PC speaker by default
Fix direct content loading of DOS executables
Fix BASSMIDI crashing during startup on 32-bit Windows
Add BASS/BASSMIDI libs to core info file
Add BASSMIDI MIDI driver –
The bass and bassmidi libraries are looked for in the frontend’s system
directory and loaded at runtime. This allows the core to work and be
distributed in a GPL-compliant way without those libraries.

DOSbox SVN

Description: DOSbox

Add SALC and XLAT to the dyn_x86 core. Improve LOCK handling a bit.
change new to new(std::nothrow) (vogons 73603) and some formatting
QNX: Adjust flags to msync based on libretro-common –
I didn’t notice any bug before but this stuff is very difficult to trigger
QNX: Add support
Allow unaligned memory only on x86
Determine CPU based on actual running platform rather than build one –
This allows to closs-compile on x86 linux into arm linux

Endianness is determined in retro_endianness.h and dynarec is determined
in dynarec.h
Correct cdrom sector size field length according to docs
memory: Add missing std::nothrow –
Given subsequent check for NUL it was obviously intended for use with
std::nothrow
Catch exceptions in dosbox –
Right now an exceptions ends up killing cothread which is against libco
recommendations and is more difficult to debug. Instead log and exit normally
Fix button mappings and wrong port assignment when using both ports –
This maps B/A/Y/X to DOS buttons 1/2/3/4 by default. Also corrects the
issue of input not working correctly when using two controllers.
Fix Windows x64 hang/crash by updating libretro-common
Don’t submit mouse emulation descriptors when no ports are connected
Fix new gamepad emulated mouse defaults being swapped
Map mouse buttons to L/R, modifiers to L2/R2, not vice-versa –
Many gamepads don’t have L2/R2, which were needed to press mouse
buttons with the gamepad using the default mapping. This change
maps mouse buttons to L/R by default. The modifiers to speed up
or slow down mouse motion are moved from L/R to L2/R2, since they
are less essential. All of this can be remapped via quick menu
using the input mapper.
Fix gamepad emulated mouse inputs not showing in mapper sometimes
Vita: Fix dynarec
Vita: Build fix
Switch to libco provided by libretro-common –
libco embedded here crashes on vita. So let’s use the common one
Use RETRO_CALLCONV for disk control callbacks –
It’s surprising that there were no crashes so far on x86 32-bit…
Fix disk_replace_image_index always reporting failure –
This silly mistake caused the “Failed to append disc” error message in
RetroArch.
Correct an oversight of r4186 when floppy disks are mounted.

Beetle PSX

Description: Sony PlayStation1 emulator

Load bios from path that was checked
sanitize gl context requests
Add more detailed error messages for mmap, unlink sooner
Allow Solaris 11 build
Cleanup warnings: snprintf truncation, strncpy->memcpy, memset
Lightrec: Cleanup/deduplicate mmap code
Lightrec: Improve homebrew support
Fix inverted check for dma-only invalidation
Fix OS X compile
Lightrec: fix more games –
These games now work:
– Need For Speed: V-Rally
– Alone In The Dark – One-Eyed Jack’s Revenge
Lightrec: Fix SWL/LWL using wrong mask
When HAVE_SHM always use global memfd so it can be closed properly on exit –
LGTM found this once the default was to compile in lightrec
Keep track of mmap failing –
If mmap failed it should prevent crashing when closing and disables
lightrec memory mirrors so as to not use an incorrect code path
Update to latest lightrec and extract PGXP from lightrec –
PGXP functions are now called from within beetle
Compile in Lightrec support by default –
Specify HAVE_LIGHTREC=0 if you don’t want it compiled
Improve DualShock calibration reference
Fix disk control interface when running single-disk PBP content

Mupen64Plus Next

Description: Nintendo 64 emulator

Only set WITH_DYNAREC based on ARCH if not set
Update GLiden64
Add support for loading GL symbols using dlsym() instead of libretro API –
* Required for platforms with EGL version < 1.5 * To enable, use new compile-time define: `-DGL_USE_DLSYM` Currently only enabled for Raspberry Pi platforms that use the legacy Broadcom driver.
Add nasm variable
Assign variable and reference nasm
Bump Version to 2.0.5
Update Mupen64Plus INI
Add fbInfoDisabled to Ini parsing
Update GLideN64 INI

Yabause

Description: Sega Saturn emulator

Enable Solaris build

Kronos

Description: Sega Saturn emulator

NOTE: This core requires you to use the ‘glcore’ video driver right now. There is no software renderer and it won’t work properly yet with the regular ‘gl’ driver.

Activate openGL program precompilation
Precompile some shaders directly at boot time to have a smoother BIOS animation
Emit a callback at each frame for synchronizing need on ports
Swap the buffers at each frame
(libretro) framerate pacing + CDROM support + m3u extension
Implement a database & rework cart auto-detect
Add support for BIOS language
Fix Assault Leynos 2 black screen
Fix mesh improved image unit usage
Fix blinking in Sega Rally, CS Mode
Add support to remove banding when using gouraud shading
Modify the handling of improved mesh handling to prepare improved banding support
fix pause in Daytona USA
fix Sega Rally USA boot – might introduce other improvements or regressions
Prepare the SH2 threading
ST-V I/O is reading words
(libretro) fix resolution mode change
Rebase the openGL on Yabause since compute CS has a better rendering and openGL rework created new issues
(libretro) improve rendering loop
Better horizontal upscaling
Depending of vertical flip, sprite reading is not the same – Improving Sega Rally
(libretro) use cpu_tesselation as default for polygon mode
Use CPU tesselation by default
(libretro) make the frame rendering more libretro-friendly
Reintroduce some required variable for color calculation – fix Cotton Club
Software renderer supports 4 threads at maximum
Consider that only old card compatible only with OpenGL 3.3 are limited in variables and need to reduce VDP2 blitting program
Reintroduce software renderer
Reduce the VDP2 register texture width
Do not initialise unsupported openGL functions. Fixing some openGL Core 3.3 errors
For openGL do not call to much the rendering loop – Better to maximize cache texture impact
Add ignition line to VDP1 commands
Do not flush the cache texture at each HBlank – only flush when needed
In case of draw every 2 frames, do not consider CMD that has already been displayed
We have to update the texture before the evaluated startupline, not at the end of the first line… Fixing Skeleton tearing
allow STV rom loading to be CRC based
Trigger the VDP1 rendering as soon as we consider the draw command list has executed
Take care of effective starting line of VDP1 command to evaluate if the core has to regenerate the textures
Remove NVidia related pragma – Might impact other GC
Calculate VDP1 cycles requested by draw commands
Fix some STV loading –
Fix batmanfr, thuntk & thunt boot without breaking dnmtdeka gfx (and maybe other’s ?)
Also got sanjeon to boot.
Load the BIOS file of the right entry – Fixing Die Hard boot
Setup EEPROM directory at the same location than STV ROM
Fix “heap-use-after-free” on exit –
YglGenReset can’t be called after _Ygl is freed
fix some STV loading issues –
suikoenb, thunt, thuntk, batmanfr & znpwfvt now boot, however only thuntk seems playable at the moment
Support EEPROM save & load for STV
Update the rotation window before the rotated layer are using the values
Preliminary support for STV’s Kick Harness
Fix horizontal RBG offset in Final Fight Revenge
Allow openGL 3.3 since openGL without tesselation should work on most of the games
Fix the RBG vertical misalignment when upscale is on – Still horizontal tearing
Some minor fixes for RBG CS
Fix VDP1 mapping on highest upscale ratio
Fixing coin setup on ST-V
Backup RAM can be accessed in word and Long – STV is doing this
Better handling of the upscale
PTRM = 0x3 corresponds to PTMR 0x2 – Fixing Skeleton Warriors startup
Fix the special color condition in case sprite is cc enabled, not active as first screen, but activable as second
For color mode other than 2, read coefficient table on the upper part of the color RAM
Do not handle prohibited setting on PTMR – Fix Skeleton
Writing lower part of Color RAM in mode 0 is overwriting upper part
When VDPRAM mode is 0, color from RBG is read on the upper part since lower part is for color offset
If the code is checking EDSR, just wait for VDP1 processing to finish – avoid changing VDP1 texture while it is changing
Do not try agressive optimization yet – Fixing Disc menu performance
Fix Quad upscale in Compute shader
Fix the regeneration of the VDP1 surfaces – Fix Guardian Heroes
Add a compilation flag to enable/disable VDP1RAM update – enabled by default
LDCSR is changing the SR mask, so interrupt shall be handled as soon as it is changed – Fixing Princess Maker 2 boot
Fix the deinit of YGLTM
Fix libretro fullscreen switch – Implement a destroy of all openGL objects when openGL context is reset
(libretro) fix upscaling weirdness
(libretro) try at fixing scaling weirdness
On Libretro, do not execute the last resizing. It is required that the libretro framebuffer always provide a FB of size _Ygl->widthx_Ygl->height
(libretro) updates
Issue on IST cleaning. Princess Maker 2 is now crashing on master SH2 issue. Might be due to SH2 interrupt handling – Fixing Skeleton boot
SH2 interrupt mechanism is not accurate – Do not try to make it precise – Fix Capcom Generations 5
Compute the VDP1 buffer in threads
Double the VDP1 structure to avoid stall between frames
Big rework of SCU Interrupt handling – Fix Princess Maker 2 boot while fixing Nanatsue Kaze after the start screen – might introduce regressions
Reenable the support of asynchronous preparation of RBG layers
Do not enable RBG async preparation
Fix BIOS HLE interrupt usage
Fix sysclip size for rotation
Check if we need to send a SSH2 interrupt each time we are sending a MSH2 interrupt from SCU
Purge all the SCU interrupt and do not stack them
Changing SCU HIRQMASK can generate an interrupt
NMI interrupt has a specific handling and does not exit using a standard RTE command
Handle in priotity interrupts on SH2 core
Rework a bit the SCU interrupt handling
Reintroduce interrupt removal – Fix Princess Maker 2 – might impact Sakura Wars
Fix Sakura Taisen video
Fix line color offset on RBG compute
Fix Highway 2000
In case of rotated FB, increase the system clipping by twice the offset – It looks like a workaround – It is fixing Hang On ’95 blue line on the right
Better to use VDP2Ram access function
Fix VDP1 rotation for Hang on and Power drift. It looks like a scale might be still needed for Capcom Generations 4

PCSX ReARMed

Description: Sony PlayStation1 emulator

Android: Support for new lightrec API
Update lightrec to latest upstream
Minimize logs when loading a cheevos-compatible content
Cleanup retro_run() –
– move input query into separate functions
– move internal fps display to separate function
Hide other inputs from core options –
– This adds a core option to hide some input options like multitaps, player ports 3-8 and analog-related fine-tuning options.
– also combine dynarec-only options in one #define directive
More core option fixes –
– This PR fixes core options and moves them to the related dynarec modes where they are implemented.

LIGHTREC = relates to platforms that supports the new Lightrec mode
NEW_DYNAREC = relates to previous dynarec implementation that is still used for some 32bit devices

– Dynarec Recompiler core option, both dynarec implementation can be enabled or disabled
Move guncon options to update_variables –
– This should stop unnecessary RETRO_ENVIRONMENT_GET_VARIABLE callback and log spamming
Fix some edge case where core can freeze upon loading content
Automatically disable Lightrec when no BIOS is present, take 2
cdriso: fix a disk switching deadlock when closing a CD image
ARM NEON: Fixed bug where MSB of a 15-bit BGR color could corrupt green value.
cdriso: fix a disk switching deadlock
unai: Add ARM-optimized lighting / blending functions

Addendum on UNAI ARM-optimized lighting/blending improvements –

“Looking at the generated ASM on 3DS, I thought I could squeeze out some extra performance by moving the inner lighting and blending functions to handwritten A32 assembly. This gives a medium improvement generally (3-5fps faster on the beach in Crash 1) and a large improvement when doing lots of blending (46-48fps before, 57-60fps after, behind the waterfall in Water Dragon Isle in Chrono Cross).

Some other notes:

I used the ARM11 MPCore (3DS CPU) timings for pipelining.
I had a few stall cycles during lighting, so I used them to preserve the MSB for lighting and blending, which saved a store, load, and orr later on. ~3-6 cycles saved overall by doing that.
I switched from u16 to uint_fast16_t, which is 32-bit on this platform. This saved a few useless uxth instructions for another few cycles. This shouldn’t affect other platforms, but I don’t know for sure. Could typedef if necessary.
A lot of the speed improvement in blending comes from not using two instructions per and. For example, & 0x8000 — the compiler preferred to mask out bytes using bic 0x7F00 and bic 0x00FF. Both slower and seemed less correct for what we’re trying to do.”

LRMAME 2003

Description: 0.72 version of LRMAME

Fix cheat input dip switch option

LRMAME 2003 Plus

Description: 0.72 version of LRMAME with updates/enhancements

See CHANGELOG for all updates/enhancements on top of LRMAME 2003

New working game Gulun.Pa! CPS1 prototype
sample pause support for journey –
uses pause instead of the mute hack implemented.
TANK III Joystick bootleg
Update samples.c
fix big samples dynamic loading when not an OST item
Update mcr2.c –
Sepways.wav sample support for journey
Update inptport.c
Update foodf.c –
Allows the player to face in the direction last applied.
A fix for Midway MCR3 game saving

LRMAME 2010

Description: 0.139 version of LRMAME

Fix Selecting “Inputs (this game)” crashes Retroarch on Android (should also affect other ARM builds)
backport 12-bit wrapping fix
Fix 12-bit wrapping behavior in YM2608/2610 ADPCM_A decoding, fixes some glitches in certain samples in the metal slug series, and likely other games. [Lord Nightmare, madbr]

Dolphin

Description: Gamecube/Wii emulator

Fix build for Windows x64
Request glcore context when video driver is gl

ParaLLEl N64

Description: Nintendo 64 emulator

Sanitize GL context requests
Update ParaLLEl RDP – about a ~10fps speed increase on Nvidia over previous version
Should fix Mega Man 64 graphics glitch (electric fence not visible)
Add ParaLLEl RDP
Add SI DMA Duration hack for Tetris 64
Use separate cache for DRAM and hidden DRAM.
Add DRAM flush and fix VI_REGISTER_OUTPUT.
Dump the hidden RDRAM as well.
Fix Seg Fault on Game Unload –
When commit 11c1ae3 split r4300_execute into r4300_execute and r4300_init, it continued to check the “stop” variable, but this is undefined. Removing these checks resolves the seg fault and does not affect functionality.

O2EM

Description: Magnavox Odyssey 2 emulator

Fix YES/NO keys and set 0 as default key
Change Action button to B to be more consistent with other cores, and remove the shortcuts to 1/2/3/4 keys (useless with the new virtual keyboard)
Add option for virtual keyboard transparency
Add graphical virtual keyboard

Opera

Description: 3DO emulator

Fix Haiku build
Remove NVRAM file and try rename again on initial failure –
Windows doesn’t like renaming over files? This logic accomidates for both
without needing platform specific behavior.

VirtualJaguar

Description: Atari Jaguar emulator

Add Haiku build

XRick

Description: Game engine implementation for Rick Dangerous

Add Haiku build

vitaQuake 2

Description: Quake 2 game engine core

Add GLES Support and initial Rockchip platform
Fix intermission screen being unskippable.
Put on par with Vita build (Bump to v.2.3).
WiiU: Add build
PSL1GHT: Add build

vitaQuake 3

Description: Game engine implementation for Quake 3: Arena

Add Haiku build

mGBA

Description: Game Boy Advance emulator

Libretro: Add cheevos support for GB/GBC
PS2: Update to newest toolchain

QuickNES

Description: 8bit Nintendo Entertainment System/Famicom emulator

PS2: Update to newest toolchain
Fix potential free(NULL); problems

FCEUmm

Description: 8bit Nintendo Entertainment System/Famicom emulator

PS2: Update to newest toolchain
user-adjustable Zapper tolerance
reduce max Zapper tolerance to 20
MMC1 overrides are treated as ines 2.0, so its needs at least default values for prgRam and chrRam columns.
Fixes FF1 pink screen due to unmapped CHRRAM.
Use proper geometry when switching NTSC filter on or off –
– Width changes previously was not respected when using NTSC filter. With full use of overscan, NES width
is 602 px when NTSC filter is used and 256 px on normal
Fix build when compiling without NTSC filter support
fixed 3DS build
Adjust Zapper tolerance; make Zapper input tolerance circular rather than rectangular by default
Fix timing when changing from PAL/Dendy to NTSC –
– Happens when starting with PAL/Dendy region and changing to NTSC can cause frame to get stuck in 50 Hz
– using RETRO_ENVIRONMENT_SET_SYSTEM_AV_INFO should remedy this.
NTSC: Remove height doubling/scanline effect –
– We just use shaders for scanline if needed. NTSC + height doubling causes performance hit
for some slow devices. Any decent platform should be able to handle scanline effect shaders at least.
ines.c: Simplify rom info logs and cleanup
Fk23c: Fix chr issues for some games –
– Affects mostly waixing using mixed chr rom/ram modes (bit 2 of ram config register $A001)
Update ines-correct.h –
– Add overrides for FK23C
– Move MMC1 overrides
– Move MMC5 overrides
Move overrides out of some mappers –
Move mapper-based overrides out and use ines-correct.h if possible. Affects the following mappers below:
– Move Cnrom database to ines-correct.h
– Mapper 201 update
– Mapper 91: Add Street Fighter III (Submapper 1) to ines-correct.h
– Add dipswitch notes to m237
– Update mapper9 (PC10 version of Mike Tyson’s Punch-out)
Move battery-backed prg ram override to ines-correct.h
Start expanding internal override database (ines-correct.h)

2048

Description: 2048 game implementation

PS2: Update to newest toolchain
PS2: Apply color correction
DOS: Add platform support

Picodrive

Description: Sega Megadrive/Genesis/32X/CD emulator

PS2: Update to newest toolchain

Snes9x 2010

Description: 16bit Super Nintendo/Famicom emulator

libretro: added granularity in SuperFX overclock –
There doesn’t seem to be any benefit of overclocking above 15 MHz
(150%). The user should be allowed to be more precise with their
overclock setting now.
Fix MMC reg for 64bit builds –
Fixes a segmentation fault when playing large ROM games.
Fix MMC bank register bit 7 (FuSoYa) –
Fixes the 64Mbit ExLoRom map.
ROM: fail if ROM is invalid –
Fixes an issue whereby a non-SNES file would cause a segmentation fault.
This may occur if the selected ROM is corrupt, or a file has an
incorrect extension.
APU: remove unused SoundSync –
Additionally modified the resampler to use buffer size as a parameter
instead of the number of samples within the buffer. Previously, the
buffer size was being changed to the number of samples, and then changed
back within the resampler.
snes9x: add defines for unused multi-cart support –
The compiler was already optimising these unused functions out. The
libretro core can define SNES_SUPPORT_MULTI_CART 1 to re-enable support
for multi-cart in the future.
snes9x: refactor defines and remove overscan –
Out of bounds memory fix from
snes9x: APU: Fix buffer overrun –
Additionally:
This fixes linking with LTO.
Disable audio if an error occurs in init instead of continuing and
segfaulting.
snes9x: reduce APU buffer to 64ms –
I believe this to be a more sane setting than a 1000ms buffer.
snes9x: backport config from upstream –
Most options were not available in English, despite it being the default
language. All the options that were available in Turkish are now
available in English.
snes9x: fix headercount increment
libretro: fix pitch measurement

Prboom

Description: Game engine core for Doom 1/2, Ultimate Doom, Final Doom

Switch from ad-hoc endianness handling to retro_endianness.h

Vecx

Description: Vectrex emulator

Fixed colour conversion (7 bit mono to RGB1555).
support analog controllers
Make line drawing code more efficient by doing RGB conversion once per line only
Nicer point shape
Allow scaling of vector display
more flexibility in adjusting display – allow for scaling and shifting to fit overlays

NeoCD

Description: SNK Neo Geo CD emulator

CD-ROM controller logic split into a separate file
Y Zooming don’t need a ROM file anymore
New system to identify and patch BIOS, should allow unknown BIOS to run.
BIOS files don’t need to have a specific name anymore they are identified by contents
Support for Universe BIOS 3.3
BIOS name in the menu now includes filename
Add synchronous CD operation mode
Vita: Use synchronous CD operation mode
Vita: Add build
3DS: Add build
WiiU: Add build
PSL1GHT: Add build
Emscripten: Add build
Implemented horizontal interrupt masking (not verified on real hardware)
Fix CDROM music endianness
Fix big-endian support

PocketCDG

Description: MP3 Karaoke audio player

Properly send logging to proper place –
use logging interface as much as possible with stderr as fallback
rather than being inconsistent
Swap frame before passing it to audio_batch_cb on bigendian –
audio_batch_cb expects native-endian frames and libmad gives little-endian
frames. Hence on big-endian we need to swap
libmad: Fix big-endian support

FreeChaF

Description: Fairchild Channel F emulator

Switch from ad-hoc endianness to retro_endianness.h
separated CHANNELF HLE code

ECWolf

Description: Wolfenstein 3D game engine core

Latest updates from upstream
iOS/TVOS: Add build
WiiU: Add build

UAE4Arm

Description: Commodore Amiga emulator

Libretro improvements
Two joystick support with automatic switching
Define controls close to PUAE default. Use right analog stick as mouse and L2-R2 as mouse button.
Add working virtual keyboard support
Emulation resolution handing is better (especially change). 2nd mouse button working.
Solve right mouse button non working
Solve different video resolution
Solve audio glitches

P-UAE

Description: Commodore Amiga emulator

Disable waiting blits by default
Remapping fixes + clarifications
Zoom mode horizontal croppings + cleanups
Prefixes for hidden core options
Better sorting in M3U generation
Geometry fixes, Keymap update, Cleanups
WHDLoad fixes –
– Custom parameter not working on many cases
– Error messages and copy skips if Kickstart filesizes are not correct
Model preset overrides, Optional region search
Core option for D-Pad joystick/mouse switching
Model config overhaul
NTSC fixes
Optional region forcing with No-Intro tag support
WHDLoad update
Sort generated M3Us, Statusbar + glue updates
Fix android build
New defaults for CPU and Drive Sound Emulation, Sound cleanups
User-friendly warning messages for Kickstarts and CAPSImg
Attempt to fix crash when reloading core on static builds
Fix ZIP Browse Archive
M3U ZIP fix
Hor+Ver positioning fixes
Add “Remove Interlace Artifacts” core option
Backport interlaced double line field mode to replace old frame mode
Fix sound filter from effectively being always Automatic at startup
Fix filter type update, Improve zoom
Add core option for CD startup delayed insert, Remove previous disc change detect trickery
Change model force hierarchy, Option label updates
Android unzippings yet again
Core option for muting floppy sound when drive is empty
Automatic horizontal centering improvement
ANDROID X86: Fix build
Fix amd64 compilation –
m68kops for amd64 is empty, use generic exactly how it was before the rewrite
Rewrite libretro m68k.h and m68kops.h to use non-libretro variants –
They encapsulate per-cpu optimizations. There is no reason to
have a version for libretro
Rewrite maccess.h using new retro_endianness
Automatic zoom improvements
Control improvements:
– Enabled D-Pad as mouse in analog joystick mode to help out menu traversing, otherwise D-Pad does nothing unless toggled to mouse mode
– Added core option for inserting RetroPads to joystick ports in different order (for Arcadia, Dyna Blaster etc.)
Fix vertical touch alignment on keyboard while zoomed
Global conf file, MultiDrive via Disk Insert
CDTV core option
Statusbar finetuning
Disk Control additions:
– Transparent ZIP support in M3U parsing
– Insert Disk support for M3U/ZIP/etc

VICE

Description: Commodore 64 home computer emulator

Keymap core option + rework
Autoloadwarp enhancement, Core option label tweaking
Remapping fixes + clarifications
Fix Plus4 cartridge launching, Statusbar fixes
Autoloadwarp fix for D81s, Statusbar cleanups
Freeze cartridge to reset types
More robust floppy autoloadwarping
VIC-20 updates:
– RAM block set tidying
– Fixed starting carts in M3Us
Achievements environment
Core option for 2nd SID, Warp mode rework
Automatic Load Warp core option
NBZ support
Dump available and not yet core optionized resources for easier ‘vicerc’ usage
Disk Control finetuning –
– Fallback to drive 8
– Remove “.” from image type detection extensions to also match cases like “.hidden-d64”
Fixed and enabled printer
Better sorting in M3U generation
Disk Control updates –
– Allow CRTs & PRGs in M3Us
– Removed redundant short pathname label fallback in widget
Automatic region fixes
Zoom mode honing –
– Changed manual mode operation from automatic to optional
– Corrected non-wide calculations
– Added VIC-II/VIC/TED border info to sublabels
Core option reorganizing –
– Separated VICE option variables from core option variables
– Prevented updating VICE variables to the same variable
– Simplified palette options
– Fixed CBM2 embedded palettes & added missing PLUS4 embedded files
– Removed & disabled non-working CBM2 models (510 not selectable in standalone, therefore no point fixing embedded files)
– Fixed Super VIC memory expansion
– Fixed CBM2 crash on higher resolution models
Work disk improvements –
– No need to reset
– Fixed start content
Automatic model core options for x64 & x64sc –
– Scans for “NTSC|(USA)” and “PAL|(Europe)” tags
– Both PAL/NTSC and C64/C64C can be preferred
– Default is “C64 PAL Automatic”
Sync finetunings –
– GetTicks returns a ticker from get_time_usec instead of fake microSecCounter
– Bypassed internal frameskips and delays
– Warp mode speed improved
– Statusbar updates FPS in 1 second intervals instead of 2 & shows real FPS from both warp and fastforward
Work disk core option with device selection
Fix for autostart detection of D71
Work disk core option with device selection
Fix QNX build
Zoom improvements –
– Presets for usual suspects (16:9, 16:10, 4:3, 5:4)
– Hardcode exceptions minimized, math maximized with science
Fix Emscripten build
Switch from adhoc endianness handling to retro_endianness.h
Sort generated M3Us
Better retro_get_region, Embedded additions
Include NIBTOOLS for automatic NIB->G64 conversion
Manual cropping core options
Remove border stuff
Zoom overhaul, Reorganizing, Cleanups
JiffyDOS & GO64 for C128
Direct hotkey for joyport switching, TDE + DSE enabled by default
Disable JiffyDOS with tapes also on static platforms
Rename mouse_x to retro_mouse_x in retrostubs.c to avoid confusion with mousedrv.c
Fix compilation on Vita
PET fixes (embedded data, keyboard layout) + Cleanups
Extended ZIP+M3U Disk Control
Fix model change not triggering geometry change with borders disabled
PSL1GHT: Add build

Frodo

Description: Commodore 64 emulator

Disable compilation of CmdPipe.cpp on libretro
Add safeguard in case of thead allocation failure
Resync libco
Add logging
3DS: Add build
PSP: Add build
Vita: Add build
QNX: Add build
QNX: Fix cmdpipe
Exclude cmdpipe on PSP1 and Vita
Don’t use sigaction on PSP1 and Vita
Don’t use chdir and getcwd on PSP1 and Vita
Change pulsehandler not to use signals on libretro

Quasi88

Fix declaration of INLINE on Android
Support for big-endian Linux
Switch from ad-hoc endianness to retro_endianness

Nestopia

Description: 8bit Nintendo Famicom/NES emulator

Fix scratchy audio in Super Mario Bros. 3 and others

Uzem

3DS: Render at half-width
PSP: Render at half-width
Android: Fix build
Gamecube: Add build
PSL1GHT: Add build
WiiU: Add build
Big-endian support
avr8: Fix initial memory cleaning –
Current code zeores-out only part of intended array
Disable av8::idle on libretro –
It’s not used by libretro
Disable recording code on libretro –
It’s not used by retroarch but it links with popen that is not
available on many of retroarch platforms
Support compilation with g++ 4.6

GME

Description: Game Music Emulator core

Gamecube: Add build
Wii: Add build
WiiU: Add build
PSP1: Add build
Vita: Add build
Use retro_endianness.h instead of ad-hoc endianness defines

Cap32 / Caprice

Description:

Fix video glitch on big-endian
Replace ad-hoc MSB_FIRST with retro_endianness.h
QNX: Fix build
PSL1GHT: Add build

CrocoDS

Description:

PSL1GHT: Add build
WiiU: Add build
Wii: Add build
Gamecube: Add build
Filter-out identical calls to SET_GEOMETRY
Fix disk-reading routines on big-endian –
Current approach of swapping after reading/before writing is sound
in theory but in current codebase it’s difficult to track and leads easily
to double-swaps. Just swap the values right before use and at assignment
Optimize byteswap
Fix big-endian video rendering
Disk endianness fix
Fix arguments to ReadPort/WritePort –
Passing 32-bit value instead of 16-bit has undefined results. It works
on some platforms and fails on other. Fix it properly
Support big-endian systems
Use retro_endianness instead of SDL_BYTEORDER
PSP: Add build
Vita: Add build

blueMSX

Description: Home computer MSX emulator

Remove leftover endiannness defines
Fix arguments for coinDeviceCreate –
Argument mismatch is fatal on emscripten
PSL1GHT: Add build
Fix non-smooth scrolling in PAL 50Hz

meowPC98

Description:

Move from adhoc endianness to retro_endianness
PSP and Vita fixes
QNX: Add build
PSL1GHT: Add build

NP2Kai

Description:

Latest updates from upstream

Bk Emulator

Description: BK-0010/0011/Terak 8510a emulator

Wii: Add build
Gamecube: Add build
Use unsigned int instead of uint –
For mingw compatibility
PSL1GHT: Add build

FUSE

Description:

Emscripten fixes

Lutro

Description:

Vita: Fix build
QNX: Add build

PX68k

Description: Sharp X68000 Emulator

Vita: Add build
QNX: Add build

Craft

Description:

Emscripten: Fix build
Fix path to auth database.
Current path ends up in current working dir which might be anywhere,
unwritable or not even exist as a concept.
Don’t attempt to run main loop after failure.
It only clouds what errors have occurred
Android: Now goes ingame and no longer crashes at startup, doesn’t render blocks yet though

Mr. Boom

Description:

QNX: Add build
Emscripten: Fix build
support big-endian linux
PSL1GHT: Add build

Xmil

Description: X Millennium Sharp X1 emulator

Android: Fix compilation
Wii: Add build
Gamecube: Add build
Add big-endian support
PSL1GHT: Add build
Add iOS arm64 build support
PSP1 scaling messes when source is larger than psp horizontal resolution,
hence render at quarter resolution
PSP: Add build
Fix QVGA support: Code for rendering at quarter of resolution is broken for libretro files
altogether as it was never really implemented.

Highres QVGA rendering forgot horizontal offset resulting in single-color lines,
fix it as well
Vita: Add build
QNX: Add build

Atari800

Description: Atari 800 emulator

Support big-endian linux
PSP: Fixes
config: don’t use unaligned access on RISC –
configure.ac enables it only on x86 and m68k. Mirror it in our manual config
Use frame counter instead of real clock

SMS Plus GX

Description:

Allow enabling or disabling FM sound (YM2413)
Prevent potential crash and cleanup –
– set max geometry height to 240
– remove some unnecessary functions and variables
– move some callbacks from retro_init to retro_load_game
Add core option (Remove Border) –
– removes left border (overscan). Works on SMS only.
– General video renderer cleanups
Add core options for region and hardware type overrides
Add proper support for PAL content –
– timing/region is detected using internal database
– fix to rom names on rom info log
Improve performance when using NTSC filters –
– Removes Sony decoder core options
– Removes doubling of height when using NTSC filters
– Use standard integers where applicable
Allow core to safely close when a required bios is not found –
– For coleco rom content, it requires the colero bios to exists. So if none is found,
just allow core to safely exit. SMS roms does not require bios to be playable, so no check is necessary for it.
Add support for colecovision roms (Experimental)

RACE

Description: Neo Geo Pocket (Color) emulator

Fix input mapping typo

TGB Dual

Description: Game Boy (Color) emulator with splitscreen link cable support

3DS: Update target

FreeIntv

Description:

Fixes graphics issues – Intellivania not working on emulator
Correctly detect MTE Test Cart
Fixes #40 – Intellicart roms w/o A8 –
Detects Intellicart roms using a different method for files that don’t begin with 0xA8
Expose RAM for Retroachievements
OSD, Keyboard keypad controls –
keys 0-9 as expected. [ and ] replace C and E
OSD updated, real messages replace cryptic “colored pixel” loading error feedback

TyrQuake

Description:

3DS: Fix build
PSL1GHT: Add build

LRMAME

Description:

Update to 0.220
add 4-way joystick simulation option
Correct 4way support

Theodore

Description:

PS2: Add build
Auto load BASIC 1 cartridge instead of BASIC 128 cartridge on TO7/70
Add virtual keyboard transparency option
Add TO8/9 keyboard
Update TO8 keyboard image and add TO7&TO7/70 virtual keyboards
Add virtual keyboards for all models
First version of on-screen virtual keyboard

gpSP

Description: Game Boy Advance emulator

Fix cheevos support

RetroArch 1.8.7 released!

May 17, 2020August 17, 2021 Libretro Team

RetroArch 1.8.7 has just been released.

Grab it here.

We will release a Cores Progress report soon going over all the core changes that have happened since the last report. It’s an exhaustive list, and especially the older consoles will receive a lot of new cores and improvements.

Highlights

There are many things this release post will not touch upon, such as all the extra cores that have been added to the various console platforms. We’ll spend some more time on that in a future Cores Progress Report post. We’ll go over some of the other highlights instead.

Netplay bugfixes

Some major netplay regressions snuck into version 1.8.5 and has remained a problem ever since. 1.8.7 finally fixes these issues.

New desktop-style Playlist View mode for MaterialUI

1.8.7 adds a new Desktop Thumbnail View to Material UI, available when using landscape display orientations. This is similar to Ozone’s playlist view. Above is a random screenshot showing what it looks like.

Notes:

The status bar at the bottom can be hidden by disabling Settings > Playlists > Show Playlist Sub-Labels
Touching/clicking the thumbnail bar toggles the fullscreen thumbnails view

This also represents a major refactor of MaterialUI’s menu entry handling code, which will make other kinds of playlist view mode easier to implement in the future.

Finally, this fixes two small existing issues:

Entry dividers now fade correctly during menu transition animations (this is subtle, and I only just realised that is wasn’t working!)
The ‘missing thumbnail’ placeholders now fade into view, just like normal thumbnails (previously, they were always displayed instantly, which was quite jarring)

Disable ‘Use Global Core Options File’ by default

1.8.7 changes the default setting of Use Global Core Options File to OFF.

This was only set to ON by default for consistency with legacy setups. There is no material benefit to this – in fact, a global core options file has the following downsides:

More file I/O – all options have to be read/written every time content is loaded or options are saved

Difficulty in editing option values by hand – e.g. sometimes this is necessary if a particular setting causes a core to crash, and if options for all cores are bundled together then sifting through them to find the one you need becomes a chore

Obsolescence – settings for old/unused/outdated cores hang around forever, and bloat the global options file without purpose. With per-core options, it is easy to remove settings for unwanted cores

Since settings are automatically imported from the legacy global file on first run when per-core files are enabled, changing the default behaviour will not harm any existing installation.

Don’t perform unnecessary cheevos initialisation when cheevos are disabled

Before, on all platforms with cheevos support rcheevos_load() is called each time content is loaded. This means the following happens even when cheevos are disabled:

If the core does not require the full content path (i.e. if RetroArch passes a data buffer directly), then a copy of the content data is made (up to 64 MB in size)
If the content is an m3u file, the file is opened and parsed to get the extension of the first file listed inside
A checksum is calculated for the content file extension
A task is pushed
A mutex is locked/unlocked several times

When Cheevos (Achievements) are disabled, all these things are unnecessary work, causing increased loading times and memory usage. On platforms with low memory (i.e. consoles) the unnecessary content data duplication is potentially harmful and may cause crashes.

1.8.7 very simply adds an ‘early out’ to rcheevos_load() which prevents the above unnecessary work when cheevos are disabled.

Cheevos: option to start a session with all achievements active

Adding an option to allow the players to start a gaming session with all achievements active (even the ones they have as unlocked on RetroAchievements.org).

When cheevos_start_active = true, instead of You have X of Y achievements unlocked, the player will see a message like this:

How the option looks in XMB:

And in Ozone:

Fallback directories for shader presets

This allows us to use the Menu Config and config file directories as fallback to store shader presets when the Video Shader directory is not writable by the user, thus following the same behavior shown by the “Save shader as” menu option.

This allows users to handle their own presets without having to mess with the directory configuration on distros such as ArchLinux, where shaders (among other assets) are managed through additional packages. But it also goes a bit further and changes the order of the preset directories, searching first on the Menu Config path, then on the Video Shader path, and finally on the directory of the config file.

This would improve the portability of the configuration for Android users, because they cannot explore the default shaders directory without rooting their devices. Moreover, I think it makes more sense, as regular configuration overrides are already being stored on the Menu Config path by default.

Testing
Assuming these directory values:

menu_config: /home/user/.config/retroarch/config/ (non-writable for testing purposes)
video_shaders: /home/user/.local/share/libretro/shaders/ (non-writable)
retroarch.cfg: /home/user/.config/retroarch/retroarch.cfg

The following menu options have all been successfully tested (see appended log output).

Save options

Save Shader Preset As

[WARN] Failed writing shader preset to /home/user/.config/retroarch/config/foobar.glslp.
[WARN] Failed writing shader preset to /home/user/.local/share/libretro/shaders/foobar.glslp.
[INFO] Saved shader preset to /home/user/.config/retroarch/foobar.glslp.

Save Global Preset

[WARN] Failed to create preset directory /home/user/.config/retroarch/config/presets/.
[WARN] Failed to create preset directory /home/user/.local/share/libretro/shaders/presets/.
[INFO] Saved shader preset to /home/user/.config/retroarch/presets/global.glslp.

Save Core Preset

[WARN] Failed to create preset directory /home/user/.config/retroarch/config/presets/Snes9x/.
[WARN] Failed to create preset directory /home/user/.local/share/libretro/shaders/presets/Snes9x/.
[INFO] Saved shader preset to /home/user/.config/retroarch/presets/Snes9x/Snes9x.glslp.

Save Content Directory Preset
[WARN] Failed to create preset directory /home/user/.config/retroarch/config/presets/Snes9x/.
[WARN] Failed to create preset directory /home/user/.local/share/libretro/shaders/presets/Snes9x/.
[INFO] Saved shader preset to /home/user/.config/retroarch/presets/Snes9x/SNES.glslp.

Save Game Preset
[WARN] Failed to create preset directory /home/user/.config/retroarch/config/presets/Snes9x/.
[WARN] Failed to create preset directory /home/user/.local/share/libretro/shaders/presets/Snes9x/.
[INFO] Saved shader preset to /home/user/.config/retroarch/presets/Snes9x/Legend of Zelda, The – A Link to the Past (USA).glslp.

Apply Changes option
[WARN] Failed writing shader preset to /home/user/.config/retroarch/config/retroarch.glslp.
[WARN] Failed writing shader preset to /home/user/.local/share/libretro/shaders/retroarch.glslp.
[INFO] Saved shader preset to /home/user/.config/retroarch/retroarch.glslp.

Run content log output
Game specific shader preset found on fallback directory
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/config/presets
[INFO] [Shaders]: preset directory: /home/user/.local/share/libretro/shaders/presets
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/presets
[INFO] [Shaders]: Specific shader preset found at /home/user/.config/retroarch/presets/Snes9x/Legend of Zelda, The – A Link to the Past (USA).glslp.
[INFO] [Shaders]: game-specific shader preset found.

Folder specific shader preset found on fallback directory
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/config/presets
[INFO] [Shaders]: preset directory: /home/user/.local/share/libretro/shaders/presets
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/presets
[INFO] [Shaders]: Specific shader preset found at /home/user/.config/retroarch/presets/Snes9x/SNES.glslp.
[INFO] [Shaders]: folder-specific shader preset found.

Core specific shader preset found on fallback directory
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/config/presets
[INFO] [Shaders]: preset directory: /home/user/.local/share/libretro/shaders/presets
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/presets
[INFO] [Shaders]: Specific shader preset found at /home/user/.config/retroarch/presets/Snes9x/Snes9x.glslp.
[INFO] [Shaders]: core-specific shader preset found.

Global shader preset found on fallback directory
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/config/presets
[INFO] [Shaders]: preset directory: /home/user/.local/share/libretro/shaders/presets
[INFO] [Shaders]: preset directory: /home/user/.config/retroarch/presets
[INFO] [Shaders]: Specific shader preset found at /home/user/.config/retroarch/presets/global.glslp.
[INFO] [Shaders]: global shader preset found.

Remove options

Remove Global Preset
[INFO] Deleted shader preset from /home/user/.config/retroarch/presets/global.glslp.

Remove Core Preset
[INFO] Deleted shader preset from /home/user/.config/retroarch/presets/Snes9x/Snes9x.glslp.

Remove Content Directory Preset
[INFO] Deleted shader preset from /home/user/.config/retroarch/presets/Snes9x/SNES.glslp.

Remove Game Preset
[INFO] Deleted shader preset from /home/user/.config/retroarch/presets/Snes9x/Legend of Zelda, The – A Link to the Past (USA).glslp.

Some other noteworthy things

RetroArch WiiU now has working graphics widgets. OSD notifications are no longer just plain yellow-colored text.
RetroArch 3DS now has basic networking and Cheevos (RetroAchievements) support.
With RetroArch 1.8.7 and overclocking, NeoCD reaches fullspeed on RetroArch PSVita. Even audio playback doesn’t stutter any more

Changelog

What you’ve read above is just a small sampling of what 1.8.6 has to offer. There might be things that we forgot to list in the changelog listed below, but here it is for your perusal regardless.

1.8.7

3DS: Add IDs for Frodo
3DS: Enable basic networking / cheevos
CHEEVOS/BUGFIX: Opening achievements list would crash RetroArch with badges enabled (on new games)
CHEEVOS: Option to start a session with all achievements active
CHEEVOS: Don’t perform unnecessary cheevos initialisation when cheevos are disabled. Should reduce startup times when loading content.
CORE OPTIONS: Disable ‘Use Global Core Options File’ by default
DOS/DJGPP: Add 32bit color support for cores
GLCORE: Switch to glcore video driver when requested by a core
LINUX/XDG: Use GenericName correctly in desktop entry
MAC/COCOA: Fix mouse cursor tracking
MENU/MATERIALUI: Add desktop-style playlist view mode
MENU/MATERIALUI/DESKTOPVIEW: When scrolling playlists, show last selected thumbnails while waiting for next entry to load
MENU/MATERIALUI: Limit tab switch rate when input repeat is active
MENU/OZONE: Fix sidebar playlist sort order when ‘Truncate Playlist Names’ is enabled
MENU/RGUI: Adjusted menu defaults, adjusted default scrolling speed
MENU/RGUI: Enable custom wallpaper when menu size is reduced at low resolutions
MENU/XMB: Limit tab switch rate when input repeat is active
NETPLAY: Fix regressions introduced in 1.8.5
RGUI: Add option to always stretch menu to fill the screen
WIIU: Enable graphics widgets