ao486 wish-list
- Caldor
- Top Contributor
- Posts: 930
- Joined: Sat Jul 25, 2020 11:20 am
- Has thanked: 112 times
- Been thanked: 111 times
Re: ao486 wish-list
For the Minimig core, it seems its possible to emulate the 68K CPU on an ARM CPU, making it work a bit like PiStorm for the Amiga.
PiStorm is a Raspberry Pi expansion for an Amiga that plugs into the CPU slot in the Amiga and then the Pi emulates the Amiga CPU. Usually the problem with emulation is that when emulating on a CPU, you cannot run every single chips instructions in parallel, which adds some possible flaws in the emulation because the timings will be off a bit. But when you only have to emulate the CPU that is much less of an issue since you wont have to deal with audio and graphics chip emulation all at once and it is much more likely to be cycle exact and have no added latency and other possible problems.
This same logic is being experimented with for the Minimig core and while its not fully working, it is working though. It boots and can run benchmarks and such... and even add MMU and FPU.
So I was thinking the same thing might be possible to do with the AO486 core and be a way for us to see a nice leap in performance, and get FPU as well. Possibly even more advanced features as well. On the Minimig core they are using QEMU, which already has x86 emulation as well.
PiStorm is a Raspberry Pi expansion for an Amiga that plugs into the CPU slot in the Amiga and then the Pi emulates the Amiga CPU. Usually the problem with emulation is that when emulating on a CPU, you cannot run every single chips instructions in parallel, which adds some possible flaws in the emulation because the timings will be off a bit. But when you only have to emulate the CPU that is much less of an issue since you wont have to deal with audio and graphics chip emulation all at once and it is much more likely to be cycle exact and have no added latency and other possible problems.
This same logic is being experimented with for the Minimig core and while its not fully working, it is working though. It boots and can run benchmarks and such... and even add MMU and FPU.
So I was thinking the same thing might be possible to do with the AO486 core and be a way for us to see a nice leap in performance, and get FPU as well. Possibly even more advanced features as well. On the Minimig core they are using QEMU, which already has x86 emulation as well.
Re: ao486 wish-list
Yeah this is the most exciting part of going hybrid. It will do wonders for the A0486. We can even imagine having a GPU running on the FPGA while the arm is running the CPU. It could probably go on to emulate a pentium 1 or 2 on the arm while you run advance features on the FPGA. And while I'm sure some people will scream "Blasphemy" most will be very happy,Caldor wrote: ↑Fri May 07, 2021 11:15 am For the Minimig core, it seems its possible to emulate the 68K CPU on an ARM CPU, making it work a bit like PiStorm for the Amiga.
PiStorm is a Raspberry Pi expansion for an Amiga that plugs into the CPU slot in the Amiga and then the Pi emulates the Amiga CPU. Usually the problem with emulation is that when emulating on a CPU, you cannot run every single chips instructions in parallel, which adds some possible flaws in the emulation because the timings will be off a bit. But when you only have to emulate the CPU that is much less of an issue since you wont have to deal with audio and graphics chip emulation all at once and it is much more likely to be cycle exact and have no added latency and other possible problems.
This same logic is being experimented with for the Minimig core and while its not fully working, it is working though. It boots and can run benchmarks and such... and even add MMU and FPU.
So I was thinking the same thing might be possible to do with the AO486 core and be a way for us to see a nice leap in performance, and get FPU as well. Possibly even more advanced features as well. On the Minimig core they are using QEMU, which already has x86 emulation as well.
-
- Posts: 75
- Joined: Wed Nov 04, 2020 10:03 am
- Has thanked: 21 times
- Been thanked: 14 times
Re: ao486 wish-list
Tempest 2000 please, and X-Quest if it's in the exodos. I tried to look at the catalogue pdf but the site was timing out - can't think why?
Re: ao486 wish-list
Before it was forced through the scaler, I could get the native resolutions on my CRT via the Analog IO board, but how clear the screen appeared was based on what speed the core was running at. I'm certain that's a major part of it.Malan wrote: ↑Thu Apr 08, 2021 4:03 pmWell, I'm not very familiar with core programing, but I'm sure he has a good reason to force it off, maybe it was dangerous for the hardware or something? I don't know. Still, I hope we could get it back one day.HMPoweredMan wrote: ↑Wed Apr 07, 2021 10:15 pmIt used to be working just fine until Sorgelig forced the core through VGA scaler.
-
- Top Contributor
- Posts: 1303
- Joined: Mon May 25, 2020 7:54 pm
- Has thanked: 416 times
- Been thanked: 399 times
Re: ao486 wish-list
It works rather well now, thanks to this custom mister file: viewtopic.php?f=13&t=2375
There is no jitter and scaling is 1:1, at least in the most common resolutions.
Re: ao486 wish-list
Hi !
Apart from some more speed I'd vote for a Gravis Ultrasound implementation. It would be fantastic for the demoscene.
An FPGA implementation of the MT-32 but as the mt32-pi is officially supported, it must mean it's kinda out of scope
Of course, I'm just dreaming aloud I'm already very happy with the current state of the core.
Apart from some more speed I'd vote for a Gravis Ultrasound implementation. It would be fantastic for the demoscene.
An FPGA implementation of the MT-32 but as the mt32-pi is officially supported, it must mean it's kinda out of scope
Of course, I'm just dreaming aloud I'm already very happy with the current state of the core.
-
- Top Contributor
- Posts: 994
- Joined: Thu Dec 10, 2020 5:44 pm
- Has thanked: 309 times
- Been thanked: 228 times
Re: ao486 wish-list
I've had problems with fluidsynth even keeping up sometimes depending on the soundfont used... would bet something timidity-based would need even more CPU. Almost like something worth offloading to the Pi. Not sure how you would communicate the GUS samples though.
Re: ao486 wish-list
another +1 for gravis ultrasound - and not just for demo scene.
I once owned one of these along with a sound blaster and LAPC-I and the gravis, in the games that supported it has:
1) interesting different set of wavetable synthesis - different to LAPC-I etc
2) higher quality sample rates IIRC
3) more digital voices (32 IIRC)
4) ability to use (and the memory management hell that was) SBOS, MegaEM and Ultramid to apply GUS sounds to non-GUS sounds.
I saw someone say adlib and MT32 covered all we needed to cover, that is so wrong from a history point of view.
http://www.gravisultrasound.com/files/d ... /GLIST.TXT
There is a reason gravis cards are quite expensive on ebay....
(now off to test my brand new misterMT32 with LAPC-I firmware and compare against the SC-55mkII I have on my desk here )
I once owned one of these along with a sound blaster and LAPC-I and the gravis, in the games that supported it has:
1) interesting different set of wavetable synthesis - different to LAPC-I etc
2) higher quality sample rates IIRC
3) more digital voices (32 IIRC)
4) ability to use (and the memory management hell that was) SBOS, MegaEM and Ultramid to apply GUS sounds to non-GUS sounds.
I saw someone say adlib and MT32 covered all we needed to cover, that is so wrong from a history point of view.
http://www.gravisultrasound.com/files/d ... /GLIST.TXT
There is a reason gravis cards are quite expensive on ebay....
(now off to test my brand new misterMT32 with LAPC-I firmware and compare against the SC-55mkII I have on my desk here )