12 Apr 2012 16:57
|I am an ex-Amiga user who sold all his Amigas and only occasionally use WinUAE - mostly to help my dad, who can't live without Amiga software. I think this is a great project with original name as well. I never liked the transition to PPC, I find every OS above V40 bloating and against original Amiga "keep-it-simple-and-efficient" design. You guys finally got the point what the real Amiga user want in the modern world.|
Anyway, I am a programmer by trade (thanks to Amiga and ZX Spectrum heritage) and would like to help you with your project if needed without any compensation. I am not a hardware wizard but I can help in software department - I know 68k ASM, GCC, AHWRM and some kernel stuff.
Have you guys been thinking to include Natami features in WinUAE to have a testbed? Though, you already have dev-boards now.
|Samuel D Crow|
12 Apr 2012 18:06
Have you looked into the AROS operating system? The new 68k versions of AROS are written in C using GCC and could use some updated AGA and ECS drivers. That would also help the NatAmi team in the future. I'm a little overwhelmed by what would be needed to create NatAmi drivers, myself, and would appreciate the help.
I don't know whether having more NatAmi team members will help matters but having more AROS team members should help a lot!
12 Apr 2012 22:45
|I am also active in the Aros-community and every help there is appreciated|
12 Apr 2012 22:57
|id second olaf. if you dont find anything here to contribute. aros(68k) needs all sorts of attention. |
homepage. archives and dev mailing list:
EXTERNAL LINK forum:
EXTERNAL LINK late commits log:
13 Apr 2012 10:21
|To answer the question about UAE, basically, no, we have not planned on it but you never know for the future.|
13 Apr 2012 21:40
|Thank you for AROS offers but I don't see myself contributing to AROS, chiefly because I am not a great fan of AROS. I respect what you guys are doing with AROS but really can't see me being part of it. I was interested in tinkering with virtual (at least for me) SuperAGA capabilities.|
In case you would like to include SuperAGA to WinUAE someday, I can help with that, I used to tinker with its sources years ago, when I wanted to port it to the PSP but I was too slow and other more talented and efficient programmers beat me to it.
|Jacek Rafal Tatko|
22 Apr 2012 18:41
|I think that including SuperAGA support into WinUAE could be practical & useful ... , and instructive .|
22 Apr 2012 21:00
|Hmm... SuperAGA is not quite specified. But you could perhaps help with that when I think about it...|
I don't know UAE so I don't know if this makes sense at all, but Natami is basically an Amiga with a much faster native chipset.
If you added various chunky modes to UAE, that could help prototype Natami. It should not be "RTG" modes, but real Amiga modes, accessible by the blitter and the rest of the chipset. I think chipmem should be 256 megs.
I think we are looking for 24-bit "byte planar chunky", + normal 8, 16 and 32 bit chunky.
You could also do a 16-bit Paula. Again, not AHI, but real Paula, only 16 bits samples instead of 8 bits. Probably also higher frequency range.
Also, I don't know if UAE can be configured to do this already, but it should have these properties to be more alike a Natami:
CPU emulated without JIT but otherwise no speed limit
fast blits and fast copper
The reason for no JIT is the Natami does not gain it's speed from the CPU, but from the chipset, so I suspect UAE with JIT will get "unbalanced" performance.
22 Apr 2012 21:36
|JIT would not unbalance performance, it is just that some loops will be rendered Useless.|
So basically using JIT you cannot run a benchmark program because it will cut out NOP loops.
23 Apr 2012 08:30
|Marcel Verdaasdonk wrote:|
JIT would not unbalance performance, it is just that some loops will be rendered Useless.
I think JIT would have the effect of varying code speed, i.e. slow on a first run, very fast on a second run through the same code.
23 Apr 2012 10:02
|Also, I don't want the CPU to be too fast in UAE. UAE JIT can be almost native speed sometimes. This would not model a Natami well. Fast memory and decent, but not super fast CPU, would model a Natami better.|
Also, if you want to experiment, you could add another m68k CPU! It is much easier to add cores to an FPGA than it is to make 1 core go faster.
Personally; I envision simple extra cores having both their own tiny local memory and access to chip mem.
23 Apr 2012 10:19
|Jacob we have talked about this before in another topic Maid cores and Robin Cores.|
IMHO what we need is something similar to the Copper's move however we need another version of wait and skip
time triggered might be better for audio operations, or event triggered like the current copper but at higher resolution, IRQ's etc...
you should know the complete list by now.
Just keep in mind KISS!
And we should just pick one and do not add another, perhaps a modes register would help wasn't there a register that only had one bit used?!?
23 Apr 2012 10:29