Castle Technology, makers of an XScale based desktop machine, announced today they had purchased RISC OS from the previous owners, Pace. RISC OS was initially developed by Acorn, starting in 1988, and has appeared in a large number of Acorn machines, and later on machines by other manufacturers.
Castle Technology Purchases RISC OS
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
RISC OS on x86 would be good. i used to use it as my first real home computer – progarmming, DTP, reports… i think the hardware is an unecessary obstacle to a very good operatinf system. i’m sure it will run fast enough under emulation on current PCs. part of the joy was the ARM assemblt language – a pleasure to code in.
t
Very good news indeed. RISC OS wasn’t exactly core to Pace in the last few years, but to Castle it most certainly is – so I’d imagine considerable development to occur (some of which may bring RISC OS to a much larger number of users).
Yep good news (there’s further on this on drobe (www.drobe.co.uk)).
Regards
Annraoi
You can get RISC OS versions to run under emulation (Virtual Acorn A5000), but these tend to run a much older version of RISC OS (Version 3.1) which has limitations that don’t exist in RO 4 or RO 5.
This may change though.
Still the fastest way to run RISC OS is on an Iyonix, and they are quite impressive when running full tilt. I’ve use Virtual Acorn on an Athlon 1.2GHz (T’bird) and it is acceptably fast, but still nowhere close to RISC OS on a real Iyonix
Regards
Annraoi
I wonder why isn’t there a nice XScale laptop, desktop or server. AFAIK NetBSD runs on it quite fine. It’s a powerful design designed for little heat dissipation/power consumption. And a dual 400MHz XScale would probably kick Transmeta’s designs’ a** in speed department because it’s Intel make and a result of better engineering (Crusoe interprets x86 in hardware). Is there any faster XScale processors that are designed for larger systems than PocketPC/Palm?
Isn’t RISCOS the bigendian OS made by MIPS to run on the Magnum PC-50 as an alternative to WINNT?
The processor used in Castle’s Iyonix (www.iyonix.com) is a 600MHz Intel IOP80321 (which is basically an xScale with an integrated DDR memory controller, PCI and other I/O stuff on the same die).
It run’s RISC OS 5, and is pretty darned quick. Also in the works is a Linux port (if you go to the iyonix link above you can get news on this).
Regards
Annraoi
Yes and no.
RiscOs is on Mips, RISC OS is on Acorn based (ARM) hardware. An unfortunate similarity of name as it happens.
The funny thing is the first OS to come out on Acorn’s early RISC machines (in 1987) was called Arthur, they probably toyed with Arthur 2 for their second OS release (but unfortunately a film with Dudley Moore called Arthur 2 came out at around the same time – so RISC OS was used instead).
I’d also add Arthur a bit “flakey” and RISC OS was much better, so a name change for less fanciful reasons may also have been warranted !
Regards
Annraoi
Looks real cool. I wish I could get the mobo+CPU only and build the case myself. It looks like they’re overcharging for hardware because there’s not much competition.
I would stick to RISC OS if my other alternative is Linux. But is there any way to install any other BSDs? Of course it runs NetBSD, isn’t it
Is there an 64 bit XScale in the works?
The price is high because they don’t sell them by the million and don’t have an economy of scale to fall back on.
Considering that the price of some “bare board” xScale development boards cost *more* than the complete Iyonix system, the price is not that over the top (mind you if they lowered the price – it might boost sales…..)
Not sure about NetBSD you could ask on Acorn/RISC OS related newsgroups such as comp.sys.acorn.misc
As to a 64bit xScale that’s down to Intel, the architecture and Instruction set it’s based on were developed initially by Acorn. Acorn eventually spun off it’s chip design section as ARM Limited (Acorn, Apple and VLSI Technology being the then owners). Many vendors (including some *very big* names produce ARM (of which some are xScale compatible chips)). One particularly tasty example is Samsungs ARM10 which should be (is ?) out soon clocked at 1.2GHz.
You can get more info on ARM processors at http://www.arm.com, when you’re reading bear in mind xScale is an implementation of ARM architecture version 5TE.
Hope that helped
Regards
Annraoi
I read about the ARM architectures and found this:
http://www.commsdesign.com/news/market_news/OEG20010731S0002
So Intel is probably going to come up with v6 architecture XScales soon. Jazelle seems to be very interesting. V6 also incorporates additional technology TrustZone as well as SIMD extensions:
http://www.arm.com/armtech/TrustZone?OpenDocument
I think that’s quite cool as well. To incorporate additional security and define access privileges with hardware support. I guess it’s not a new idea (didn’t Sun have similar hardware security features?) and could be used for customer control purposes (like Palladium). But at the same time enforcing user/process access privileges with hardware help on a UNIX system appeals to me.
So XScale will also move into desktop, and so will IA-64 as x86 fades out. Intel would get quite strong in server market with the rise of Itanium ( http://news.com.com/2100-1006_3-1023146.html?tag=lh ). EPIC is an amazing technology, and coupled with this immense investment it’s bound to succeed. On embedded devices Intel XScale will dominate the market.
That’s quite good news. As x86-64 attempts end up in failure I for one would be very happy to get rid of decrepit x86 architecture. Backwards compatibility? who cares? Who runs 16 bit applications today?
(The guy who submitted the item and did the Iyonix Linux port)
A 64-bit XScale is exceptionally unlikely any time soon. Intel have no market for it, and it would mean loads of architecture changes.
The price of the machine reflects the substantial costs involved in developing a machine (and its OS) for a very small market. There’s no monopolistic practices here. There may be a motherboard only option in future with only a Linux bootloader on, although I won’t discuss the price of that here.
A NetBSD port is in the works, but I do not know the current state of play.
RISC OS on x86 makes little sense (despite apparent, but flawed argument about price) business wise, and gains you much less technically than you’d lose. ARM hardware’s biggest win is in power consumption for embedded devices.
There isn’t an XScale laptop, because there are endless challenges in getting in supplying a case for what would be such a tiny potential market (probably only hundreds).
Depends on how much memory you want to support. My guess is 5 years so you can keep a full lenght movie on your handheld
I’ve deduced you’re replying to the 64-bit point, since you’ve not stated this.
It’s not a requirement of the processor nor the OS to be 64-bit just to read 2 (or 4) GB files. The Linux filesystem layer demonstrates this.
XScale does not have a floating point arithmetic unit. This will show up as inefficient (fpu emulation) or poor quality audio/video decoding.
For example, mp3 decoders written for XScale/StrongARM typically uses integer math with the quality yielded by the decoding algorithm being below that specified by the ‘mp3 standards’.
I can’t say whether it’s XScale, but – http://www.riscstation.co.uk/html/portable.html offer a RISC OS Laptop, or at least will be offering one.
They don’t currently list the processor specification, but if you’re interested in a RISC Laptop it certainly looks interesting.
Don’t expect that laptop to ever appear…
For example, mp3 decoders written for XScale/StrongARM typically uses integer math with the quality yielded by the decoding algorithm being below that specified by the ‘mp3 standards’.
Not true. MAD decoder, one of the best MP3 decoder out there is integer based. There’s also Ogg Vorbis Tremor, for integer based ogg vorbis decoder. AFAIK You don’t lose computation power by restricting yourself to integer arithmetic. Neither do the processors do complex arithmetic but we can always achieve the same result. Integer vs. FP is analogous….
With Castle, who have their own, home-rolled RISCOS 5, what will happen with RISCOS, Ltd. and their RISCOS 4? Will they be able to keep their OS licence, or will they perhaps be charged with the further development of a unified RISCOS, so that Castle may concentrate on the hardware?
Seems like business as usual. Have a look at http://www.drobe.co.uk and http://www.iconbar.com. Paul Middleton (ROL’s MD) seems happy enough.
I’d imagine RO4 will continue to be used by RiscStation and Omega (when/if ever it finally appears). As to the “bells and whistles” version of RO 4 (Select Issue 3) is still due out. It also appears that some “Select Features” may be added to Castle’s RO5. I would not be surprised if RISC OS Limited will still get license fees from RO4 sales to RiscStation and Microdigital (Omega) and pass on Pace’s (RISC OS’s former owner) cut onto Castle (RISC OS’s new owner).
Longer term RO 4’s future I would say is limited (as it can’t be run on 32bit Only faster ARM9/10/xScale processors). The existing users (however) will still be able to avail of progressive upgrades in the form of Select.
Regards
Annraoi
MP3 Decoders: No, that’s not correct. In fact, the RISC OS MP3 player (which makes use of fixed-point arithmetic) is one of the best around. There’s nothing “below standard” about it. I am also realibly informed that “there’s no quality lost by using integers, in many cases it’s the complete opposite”.
The RiscStation laptop is both woefully underpowered (56MHz 7500FE processor) for 2003, and dreadfully late (2 years and counting).