Is Yabasanshiro emulator coming on retropie?

barbudreadmon

@Darksavior said in Is Yabasanshiro emulator coming on retropie?:

@myzar It was set to interpreter by default. I changed it to dynarec and speed now matches the 64bit version.

Yeah, as i previously said, the gpu is probably the bottleneck here, yabasanshiro shouldn't be cpu-limited on a72@2Ghz (it runs fine on a57@1.43Ghz), it might be a different story at stock clock though (?). Did you overclock your gpu already ?

grant2258

@roslof said in Is Yabasanshiro emulator coming on retropie?:

By not providing "FindQt5WebEngineWidgets.cmake" in CMAKE_MODULE_PATH this
project has asked CMake to find a package configuration file provided by
"Qt5WebEngineWidgets", but CMake did not find one.

sudo apt install qtwebengine5-dev

Darksavior

@barbudreadmon No, this was just a brief test. I just wanted to see if there was a significant difference between the 32bit and 64bit versions of the non-libretro version. I really needed to test a game with a constant fps to know for sure, but the difference in sega rally was unnoticeable.

roslof

@grant2258 said in Is Yabasanshiro emulator coming on retropie?:

@roslof said in Is Yabasanshiro emulator coming on retropie?:

By not providing "FindQt5WebEngineWidgets.cmake" in CMAKE_MODULE_PATH this
project has asked CMake to find a package configuration file provided by
"Qt5WebEngineWidgets", but CMake did not find one.

sudo apt install qtwebengine5-dev

Appreciated man, as always. Probably not going to work for me on the Pi4:

pi@retropie:~/build $ sudo apt install qtwebengine5-dev
Reading package lists... Done
Building dependency tree
Reading state information... Done
E: Unable to locate package qtwebengine5-dev

But seriously, appreciated. Thanks for tryin'! Using the compiled version posted earlier.
-Ros

barbudreadmon

@Darksavior said in Is Yabasanshiro emulator coming on retropie?:

I just wanted to see if there was a significant difference between the 32bit and 64bit versions of the non-libretro version

Apart from load and temp which might be slightly higher on 32bits, there shouldn't be a difference with the cpu at 2Ghz. Also, fwiw, on the libretro port there is a bottleneck if you have vsync enabled in RA config (as a matter of fact standalone disable it).

shavecat

Hey,
Is it Worth the install ?
if so how do i do that :)? thanks

NecFan

Not sure that's really the right topic, but I just read the 247 messages and it looks having a decent Saturn emulation interest many people there : p
2, maybe 3 months ago, Shima, SSF's autor, bought a Raspberry Pi 4 to try to port his emulator himself.

Please, look at this message and others around:

https://twitter.com/FessX/status/1242075475414937600?s=20

Maybe Shima could be successful with some help from this board ?
SSF is notoriously the best Saturn emulator.

barbudreadmon

@NecFan said in Is Yabasanshiro emulator coming on retropie?:

SSF is notoriously the best Saturn emulator.

There are many criterias for "best emulator" : speed, compatibility, enhanced features (like improving original gfx), ...
SSF is well-balanced : it's fast (but yabasanshiro might be even faster with its dynarecs & gpu acceleration ?), it's compatible (iirc ~92%, only second behind mednafen/beetle at ~95%), but afaik it doesn't have any enhanced feature.

On a raspberry pi i think it's a really good prospect, but i wouldn't call it the"best Saturn Emulator" since imho there are superior choices if you aren't playing on a toaster.

grant2258

sum vulkan news https://www.raspberrypi.org/blog/vulkan-update-now-with-added-source-code/

Join us on irc.freenode.net, #videocore channel. For specific questions like the issue we have here

steeeb

@grant2258 has this news been added to every thread or something? I've had 4 notifications from it just from threads I've been involved in, and I've seen it in at least 2 others!

Everyone is probably already aware from the existing topic on it.

grant2258

this was posted 3 days ago cant control other threads where else is it posted? what you point here did you post the same question in every other thread it in. Im a little confused on your picking this thread alone.

barbudreadmon

@steeeb yabasanshiro doesn't support vulkan anyway... furthermore i think fixing gles support should be a priority instead of adding support for another gfx api...

grant2258

@barbudreadmon said in Is Yabasanshiro emulator coming on retropie?:

@steeeb yabasanshiro doesn't support vulkan anyway... furthermore i think fixing gles support should be a priority instead of adding support for another gfx api...

Vukan is the way forward it wil be the defacto standard in time

barbudreadmon

@grant2258 said in Is Yabasanshiro emulator coming on retropie?:

Vukan is the way forward it wil be the defacto standard in time

While i agree with that, support still has to be added to the emulators, and fixing current issues should be a priority over implementing new stuff.

grant2258

@barbudreadmon said in Is Yabasanshiro emulator coming on retropie?:

While i agree with that, support still has to be added to the emulators, and fixing current issues should be a priority over implementing new stuff.

Its been said before to get this resolved you need to identify the issue and report it in a way that is meaningful to devs.

Its the same when A bug appears in libretro users are asked to provide enough information to work with. If the devs know the specifics they can work on fixing it. Is there more than one emulator suffering from this issue? afaik its just here i could be wrong.

barbudreadmon

@grant2258 said in Is Yabasanshiro emulator coming on retropie?:

Is there more than one emulator suffering from this issue?

No idea, the only thing i know is that the pi4 is the only arm+gles3 device having this issue with yabasanshiro.

Can you define meaningful report ? Is reporting how to reproduce this issue on their forum not sufficient ? I consider myself lucky when i get that much in a bug report.

grant2258

@barbudreadmon said in Is Yabasanshiro emulator coming on retropie?:

No idea, the only thing i know is that the pi4 is the only arm+gles3 device having this issue with yabasanshiro.
Can you define meaningful report ? Is reporting how to reproduce this issue on their forum not sufficient ? I consider myself lucky when i get that much in a bug report.

No not really your asking a dev to dig through a whole emulator. Your not singling out or narrowing the problem.

You cant tell them where it fails what layer is causing it and what function is only a real dev with real experience with this core and gles can report a bug effectively.

Its like saying look linux crashes with this app a fix it. The dev from the app awould need to identify the suspected cause and report it to gles team where its failing. Asking gles folks to work out if its your apps issue or there themselves is a big ask. Im sure someone with more experience will report the issue one day in a way that will get it fixed. It would actually be good to have more fails else where to help narrow the issue down in all honesty

barbudreadmon

@grant2258 i think https://www.raspberrypi.org/forums/viewtopic.php?f=68&t=270359&p=1641195&hilit=yabasanshiro#p1641195 is already providing some nice information though, i would be very happy if i ever had a bug report with that much details in any project i'm involved with...

grant2258

I dont see any definitive info there. It could still be a core or driver bug at this point. Its a start though anyway dont want a 1000 posts on this again. A solid report with concrete info will sort this issue wherever it is.

https://github.com/libretro/RetroArch/issues/10711 was assumed to be gcc 10 an it wasnt the truth is stranger than fiction sometimes

barbudreadmon

@grant2258 said in Is Yabasanshiro emulator coming on retropie?:

https://github.com/libretro/RetroArch/issues/10711 was assumed to be gcc 10 an it wasnt the truth is stranger than fiction sometimes

More recent version of compilers being sometimes more sensitive to program bug is nothing new, and the whole bug looks like something that would have been totally avoided from the beginning whatever the compiler if the coder used something like ASAN. This bug can't be compared to a bug where one program work properly on every known gles3 devices minus one.