The next PowerDev Meeting has been scheduled. It will be held 28 – 29 March in Lublin, Poland. The strong organizing team should be as successful with this event as the last one they planned. In addition to the items covered on the Meeting webpage, the rumor is that the Mac mini (PowerPC) MorphOS port will be demonstrated.
In the meanwhile, the first i.MX515 PowerDev Kits
should be on the way to selected Developers.
The first samples of the new PowerDev Kit arrived this past week. The touchpad rests on top of the display. We will have many more soon. We may not have them shipping in time for the PowerDev Meeting in Lublin, but we will get them out as fast as we can. It would be great to see an i.MX515 MorphOS port.
If you have not registered your potential project in the i.MX515 Developer Program, please do so. There is change afoot!
The Community is the Computer - a Super Computer. Go Zig!
12 comments:
I would like to thank Genesi for great support of this incoming PowerDev Meeting
as well as promotion of AmiParty and other Amiga meetings which took place in the past.
I'm very glad that I could send reports from these successful events and read them on this blog then.
I hope we will co-organize more PowerDev Meetings and other projects with Genesi in the future
Best Regards,
Jerzy "Drako^lM" Guc
It's good to see you still working on something that could be useful for mass production, something that might interest millions of consumers.
The fact that the new board uses an ARM instead of a PPC is a bit disappointing for me. But this is a result of the changes ongoing on the CPU market.
As an MophOS fan I would like to read some encouraging comments of the MOS team, a port of MorphOS might be possible because the endianess of the ARM MCU should be right for a port and it would demonstrate the flexibility of the OS.
But I don't expect any announcements here and now.
cu GK
Well, this first appeareance of an ARM based computer in a POWER meeting should help a lot in its acceptance...
Like Gerd I also would like to hear some word by the MorphOS team about ARM. I also think that due to no endianess problems on ARM a MorphOS port actually *is* feasible. Of course no Sunday afternoon walk, but a lot of work. Biggest job may be a good 68k emulation. Also the question is whether a hypothetical MorphOS on ARM should include a ppc emulation or not.
/U.
@ujb
There might be a problem in that m68k and PowerPC are traditionally big-endian and ARM is little-endian. Natively compiled applications simply will not notice the difference however unless they figure on loading data from disk or network which is big-endian and never bothered to swizzle it.
There's also the very simple architectural problem that m68k and PowerPC processors have more registers available - PowerPC took a few hints and made special registers available for many things like function linkage (LR) and the program counter (PC).
ARM is proper RISC - it has 16 general purpose registers the vast majority of which are used to hold things like the program counter, return address, stack pointers. This means you get a lot less to work with when you are in code.. and performance and some things like making system calls (from "user" to "supervisor" mode) where you can't carry through the caller's stack.
You could make a very decent case against m68k emulation; how much m68k software does anyone really use in MorphOS right now? Is it really ingrained in the architecture of the OS to have this virtual m68k floating around, to pass old library calls (anything that doesn't have a PPC Native Jump Table..) through this pretend processor, even if it's native PPC code behind?
How much software for MorphOS really relies on PowerPC? Legacy code from PowerUP or WarpUP, perhaps. You would not see those old copies of Wipeout running on Goa.. but is that a huge loss? How much of it is open-source or has an active developer who just needs to install an ARM toolchain and type "make"?
Sputnik and OWB would get instant benefits from being able to take advantage of the ARM optimizations in WebKit.
A very large amount of code running on MorphOS is just a Linux port anyway (usually on top of SDL).
And the rest? You may lose AmiNetRadio because it's written in AmigaE, which took a long while to get the PowerPC native ECX compiler. Would you wait for an ARM version of E or would you just run out and find yourself a different internet radio streaming music player?
@ Neko
But as far as I know the ARM can do both endiannesses: Normal mode is LE, but it can also be operated in BE mode. Does BE mode on ARM yield to some disadvantage?
True, 68k emulation is nice, but indeed not always needed. The most important 68k binary today used on MorphOS is most likely Arexx (I wonder if the reimplementation will ever get finished...).
Personally I use a couple of 68k apps, but only very few on a regular basis.
Can anyone recommend a good ARM community site?
>But as far as I know the ARM can
>do both endiannesses: Normal mode
>is LE, but it can also be operated
>in BE mode. Does BE mode on ARM
>yield to some disadvantage?
AFAIK some ARMs could do both andianesses.
IMR a biendieness doesn't means it will emulate the other endianess perfectly ...
Can you guys point me in the direction of a good OpenPVR project that features an ARM or PPC based board with ATSC/QAM tv tuner & SATA controller on board?
@ujb
Big-endian support on ARM is the same as little-endian support on PPC - it's not so much support as it fudges certain things to load in little endian data (the registers are still big endian mode).
Instruction format is still little endian, addressing modes are still little endian.. it's not comprehensive at all.
@anonymous
There's no such thing, not since we discontinued the HMC.
I hope to see PegasosIII instead. I think it will be sad if MorphOS goes ARM. But then all 68k compatibility would be gone. I hope for a PegasosIII machine instead and macmini G4 port of MorphOS
Post a Comment