Uae4arm: Amiga emulator with JIT and DispmanX

[Haemogoblin]

Will be swinging by my GFs later, where my A600pi resides. Will bring it back with me and map the keyboard.

[pawelkrak]

Fresh compilation with n0rt0nthec4t new led code.
GUI entry key changed to F10 becouse CTRL + Amiga is used to reset UAE4ARM

http://www.filedropper.com/uae4arm_2

[n0rt0nthec4t]

I found some trick:

setleds -L +caps< /dev/tty1 that work from console and lit Amiga CAPS LOCK on A1200 kybrd.

Yep, i know that lights the led, but what i dont have working is if you press the capslock key, the led should come on.. If the key is pressed, it should go off.. doesnt seem to work

[skn]

Hi Chips,

I've just found out the latest uae4arm (v0.5) has an issue with the audio (it runs slower when compared to v0.4 for some games - Dyna Blaster for example). Probably due to the new JIT code?

Thank you for uae4arm! Keep up the great job!

[Haemogoblin]

Ok took me a little longer, but I have my A600Pi setup. I'm typing on it right now!

I will sort out a keymap later this evening, Imagine its going to take me a while to do, as the only way I know to do a keymap is to use the showkey command and press each individual key.

[Haemogoblin]

I just downloaded and tested this file on my A600PI. After a quick lap on Stunt Car, all I can say is you have me deepest thanks! Worked like a charm! All the thrills and spills of the latest UAE4ARM, minus the debilitating keyboard problems! Engaging cheesy grin mode

Fresh compilation with n0rt0nthec4t new led code.
GUI entry key changed to F10 becouse CTRL + Amiga is used to reset UAE4ARM

http://www.filedropper.com/uae4arm_2

[pawelkrak]

Glad to heard this but remember its not my work. Most hard work do @Chips.

Have a nice gaming with gold years of Amiga. Have fun .

[Chips]

I've just found out the latest uae4arm (v0.5) has an issue with the audio (it runs slower when compared to v0.4 for some games - Dyna Blaster for example). Probably due to the new JIT code?

The JIT code hasn't been updated lately...
I changed some semaphore management in audio part so that we use more buffer... this should improve audio and not make it worst... but we never know

https://github.com/Chips-fr/uae4arm-rpi ... 27825e039a

I will look if i have same result on my setup with dynablaster...

[Haemogoblin]

Glad to heard this but remember its not my work. Most hard work do @Chips.

Have a nice gaming with gold years of Amiga. Have fun .

Chips HAS done an amazing job with UAE4ARM and my hat goes off to him for all the time and effort he's put in. Thanks to you though, those of us with Keyrah interfaces and no F12, can still appreciate the most recent updates!

[skn]

Hi Chips,

Thank you so much for the feedback. I'll revert the code to the original one and compile it to verify if that solves the audio issue. I'll keep you informed!

[PhillyFromNL]

I've just found out the latest uae4arm (v0.5) has an issue with the audio (it runs slower when compared to v0.4 for some games - Dyna Blaster for example). Probably due to the new JIT code?

The JIT code hasn't been updated lately...
I changed some semaphore management in audio part so that we use more buffer... this should improve audio and not make it worst... but we never know

https://github.com/Chips-fr/uae4arm-rpi ... 27825e039a

I will look if i have same result on my setup with dynablaster...

Hi Chips,

Isn't it a idea to dedicade one of the RPI core's just for the audio thread?

Gr,
Phil.

[skn]

Hi Chips,

Thank you so much for the feedback. I'll revert the code to the original one and compile it to verify if that solves the audio issue. I'll keep you informed!

Hi Chips,

It didn't work... Even after removing the semaphore management code, the audio is still slow when compared to v0.4... It must be something else...

Thank you!!!

[n0rt0nthec4t]

Will be swinging by my GFs later, where my A600pi resides. Will bring it back with me and map the keyboard.

Thanks. That would be great. Do you run key rah with the switch in middle or down?

[Chirone72]

The JIT code hasn't been updated lately...
I changed some semaphore management in audio part so that we use more buffer... this should improve audio and not make it worst... but we never know

https://github.com/Chips-fr/uae4arm-rpi ... 27825e039a

I will look if i have same result on my setup with dynablaster...

Hi Chips, thank you for your great work, I am one of the admin of FB group 'Retro Emulation Pi - FPGA & Co.', we are producing some "pure" images, completely free, that starts directly with emulators to spread these new modalities to make retro computing. Last image was done with minibian from our member GC and, from yesterday, I'm deeply testing it and I've encountered these issues:

Benchmarks in 0.5 performs worst than in 0.4 (Rpi3 and SYSINFO V4.0)
Dhrystones without JIT 12798 (0.4) now is 10542 (0.5)
Dhrystones with JITin was 74057 (0.4) now is 61015 (0.5)

Not happy about these results I've tried several adf tested in the past and there is often something that is not clear, then I've tested Forgotten World, although we could easily do without this game, it lags in 0.5 but it works enough smooth with 0.4 (tested is several different configurations/machines).

Hope that also this reporting could help you to solve recent issues

Many thanks for your contribute with this fantastic emulator !!!

[Haemogoblin]

Will be swinging by my GFs later, where my A600pi resides. Will bring it back with me and map the keyboard.

Thanks. That would be great. Do you run key rah with the switch in middle or down?

I run it without a switch at all :-p as I removed it. But I think it was in the down position lol

[n0rt0nthec4t]

Fresh compilation with n0rt0nthec4t new led code.
GUI entry key changed to F10 becouse CTRL + Amiga is used to reset UAE4ARM

http://www.filedropper.com/uae4arm_2

Seems that the 4.4.x kernel series might have a bug that doesn't light the scroll lock led when required (testing rom cmd line using settled -L +scroll) also fails.. *sigh*

Update: To get the scroll led work on 4.4.x kernel (raspbian in my case), need to create a udev rule in /etc/udev/rules.d Called mine 50-leds.rules with the following entry

ACTION=="add", SUBSYSTEM=="leds", ENV{DEVPATH}=="/input::scrolllock", ATTR{trigger}="kbd-scrollock"

[n0rt0nthec4t]

I've been playing with keyboard leds some more.. This attached source file is based off the idea neurorulez raised in the github issues log. essentially, the caps lock key doesn't work setting the led status.. Testing his idea, showed that the specific debian workaround code to handle capslock setting didn't work as expected.. need to press caps lock multiple times for it to other turn on or off.

The attached changes to pandora.cpp mirror the state of caps lock on the keyboard when uae4arm is started (so if caps lock is one, caps lock in the emulator is on). Also setting the keyboard leds to match

Happy for people to test and see what you think. Link to modified file (based upon 0.5 of CHIPS code)

http://s000.tinyupload.com/download.php ... 4367015186

I've been testing this on a raspbian (2016-05-27) release (kernel 4.4.11)

[HoraceAndTheSpider]

still a shame to hear no update on improved control mapping (just a "this is / isnt being worked on" would be appreciated)...


On another matter though, is it possible to compile UAE4ARM for the pi, so that if launched as ./uae4arm gamename.uae , then the config is loaded with launch?

I have tried doing this by changing the ES_Emulator.cfg to register '.uae' files as Amiga "roms" , and using this to launch individual games, however, the config seems to be ignored on launch?


thanks

[n0rt0nthec4t]

On another matter though, is it possible to compile UAE4ARM for the pi, so that if launched as ./uae4arm gamename.uae , then the config is loaded with launch?

./uae4arm -f gamename.uae


that would assume your saved uae config has DF0 configured with the appropriate add etc

[HoraceAndTheSpider]

Thank you!

I will be using HD mounts, but same principle!

On another matter though, is it possible to compile UAE4ARM for the pi, so that if launched as ./uae4arm gamename.uae , then the config is loaded with launch?

./uae4arm -f gamename.uae


that would assume your saved uae config has DF0 configured with the appropriate add etc

[Chips]

I did try to merge custom control.. then second joystick wasn't working any-more. Did solve the later problem but there could still have some others impacts since the code is quite messy...
The whole panel needs a refresh though...

[HoraceAndTheSpider]

I did try to merge custom control.. then second joystick wasn't working any-more. Did solve the later problem but there could still have some others impacts since the code is quite messy...
The whole panel needs a refresh though...

Would the PSPUAE source be any use to you? I can supply it if needed.

It has mapping of cd32 controls and also mouse to joypad routines...

[Chips]

This is not my priority. What do you mean by mapping CD32 controls ? Wheren't they normal DB9 joystick ?

[HoraceAndTheSpider]

This is not my priority. What do you mean by mapping CD32 controls ? Wheren't they normal DB9 joystick ?

They were but you cannot currently map cd32 buttons (only a selection of joystick bits and keyboard buttons) to your own joypad on uae4arm (at least not on the pi versions I've tried)

Mouse also was db9, but it think cd32 used a similar system to the mouse analogue data, as there are more cd32 buttons than signal wires.

I appreciate not a priority, but hopefully not too tricky (for cd32 at least) as many whdload games map keyboard controls to cd32 buttons, it would make many game much more playable on a practical level, and may already be sitting there in the code just needing adding to the "options" list of available maps

[Chips]

For those who complains about F11 & F12 not available in their keyboard, I add a configuration in panel input.
So now you can set key to go back to menu with any of these keys: F11 , F12, Left ALT , Left CTRL.
If more key selection is needed let me know. This config can be even saved (and yes you can load automatically a config)

I've just found out the latest uae4arm (v0.5) has an issue with the audio (it runs slower when compared to v0.4 for some games - Dyna Blaster for example). Probably due to the new JIT code?

Dyna blaster works well on my Rpi 3...

Not happy about these results I've tried several adf tested in the past and there is often something that is not clear, then I've tested Forgotten World, although we could easily do without this game, it lags in 0.5 but it works enough smooth with 0.4 (tested is several different configurations/machines).

At some time a new jit code was used which offer better compatibility at a lower speed... there is always a trade off to find and I don't see this as an issue as of now.
I tried forgotten world and this one definitely has some strange behaviour in latest uae4arm, but it looks like more something related to wrong internal emulation timings than lagging due to missing of CPU power from the Rpi...