RISC OS news and information from Drobe


Beta! \| About us \| Contact \| RSS	Webspace \| Tech docs \| Downloads \| BBC Micro \| Gallery \| Wallpaper

StrongARM last orders passes

By Chris Williams. Published: 24th Jun 2003, 22:05:42.

We'll miss you SA110

Stop the clock! Drobe reader "Sprow" has this evening pointed out that the 'last orders' deadline on the SA110 aka the StrongARM has passed. According to an Intel Product Change Notification [pdf], the StrongARM processor was discontinued during December 2002 however developers have been given until the 16th June 2003 to place "last time buy" orders for the chip.

Employing the Harvard Architecture and boasting two caches and speeds over 200MHz, the StrongARM started off life as a joint venture between Acorn and Digital. The first prototype was powered up in March 1996, according to Robert McMordie's Acorn History. The release of the StrongARM processor card and RISC OS 3.7 left the RISC OS userbase on quite a high in 1997.

Everyone will agree that the StrongARM has served us all very well and you may recall a 300MHz version appearing in the Castle Kinetic range at one point. But it's 2003 now and there's newer, faster alternatives to the legacy StrongARM processor. The Intel PCN note clearly encourages customers to "evaluate other Intel XScale microarchitecture-based products", adding that "possible replacements for the SA110 are the IOP310 chipset, the IOP321, and the 80200, as well as other processors soon to be launched."

So the IOP321 processor used in the Castle Iyonix is safe for now, however we don't know how this final discontinuation of the StrongARM processor affects the MicroDigital Omega. The Omega employs a 300Mhz StrongARM chip but MicroDigital expect to follow this up with an XScale second processor solution. The discontinuation will however shut the door on any future proposed StrongARM based designs.

So raise a glass to the SA. Final SA110 orders placed in time before the 16th June deadline are expected to ship before the end of 2003.

Links
Intel

Discussion

Viewing threaded comments | View comments unthreaded, listed by date

-Sob-

Anyway, pulling myself back together.

Two thoughts, once the SA-110 finally goes the fastest processor able to run RISC OS (26bit) will be the venerable ARM7500FPE (isn't the latest version 64MHz ?).

It also (and this *is* truely wierd) means that the Omega (if it takes much longer) won't have it's primary processor (xScale doesn't run on it yet, and wasn't the SA-110 supposed to run the 26 bit code - or is how ARMTwister works changed (again ?))

Anyway MD can use an excuse like "Sorry gov' we got the computer done but then Intel goes and bloody kills the processor orf" or something to that effect....

Where are Gilbert & Sullivan when you need them ?????

-- Annraoi McShane,

AMS on 24/6/03 11:39PM [ Reply | Report ]

I blame the press -- Spriteman.

Spriteman on 25/6/03 12:14AM [ Reply | Report ]

I remember an interview somewhere ages ago where this was raised with MD, and they replied something along the lines of "You think we haven't taken the appropriate precautions?"

moss on 25/6/03 12:17AM [ Reply | Report ]

I hope MD ordered enough. ;o)

(unless they're using the SA1110 or something similar) -- Simon Wilson, Boulder, Colorado

ksattic on 25/6/03 12:27AM [ Reply | Report ]

Just to clarify: when I say "this was raised", I mean that the possibility the chip might cease production fairly soon was raised - not this particular announcement, and not confirmation that the chip would stop production. The interview was about a year or so ago, I think.

Yes, that's my version of a clarification

moss on 25/6/03 12:33AM [ Reply | Report ]

The StrongARM is dead...Long live the StrongARM...More than any other single chip I feel the StrongARM has had more of an effect the world computer market. It will forever be regarded in good stance!!! :-)

Just out of idle curiosity had anyone tried commisioning ARM Ltd. To desgin a chip ideally suited to running RISC OS at high speed - after all the SA was put together in next to no time - perhaps it's time for a 'True' StrongARM 2?? As ever RISC OS will survive whatever the market throws, but perhaps it's about time ARM Ltd remembered it's origins and paid back the favour, who knows perhaps the day of an ARM/RISC OS global standard may still be achieved - hey I'm an eternal optimist, allow me my dreams :-)

So come on ARM, sod the portable market and take a chance on throwing Wintel out of the desktop market too!! :-)

-- Ryan Hitch, Cottingham

drjones69 on 25/6/03 2:55AM [ Reply | Report ]

I hope you are kidding. Such a prospect would cost a fortune, and tying future RISC OS machines to yet more obscure hardware would be a disaster. And in any case, we already have an excellent faster processor for running RISC OS on - the IOP321. Found, oddly enough, in the Iyonix. We really don't need to perpetuate the 26-bit support.

-- Peter, drobe.co.uk

mrchocky on 25/6/03 6:32AM [ Reply | Report ]

This is a sad day. The end of a very successful processor. Ah well, we've got XScales now

Does this mean the Riscstation laptop really is dead? There was some talk a while a go it was going to be SA-110 based. And I think their mythical desktop machine (Evolution) was supposed to have this processor too.

timephoenix on 25/6/03 7:17AM [ Reply | Report ]

Can something be dead if it was never alive?

thegman on 25/6/03 9:52AM [ Reply | Report ]

thegman - I sure there must be a Buffy episode which would explain that. :-)

quatermass on 25/6/03 1:46PM [ Reply | Report ]

I thought the RiscStation laptop was going to be 7500FE based? -- Ian Hawkins (g0tai)

piemmm on 25/6/03 1:56PM [ Reply | Report ]

quatermass: I don't normally watch Buffy, but those episodes I have seen were certainly based on that premise.

g0tai: There was some mutterings a while ago about them thinking about making an SA laptop, I guess they thought demand for a 64Mhz laptop for over a grand would be a little limited.

thegman on 25/6/03 2:23PM [ Reply | Report ]

mrchocky> Shame that the Iyonix only runs a s***ty old version of RISC OS with some bastardisation at the bottom end to make it run on the XScale. Until it runs something decent, you can't say an Iyonix is a good replacement for a RISC PC.

imj on 25/6/03 6:40PM [ Reply | Report ]

Ooh look at g0at's "s***ty" filter in full effect there. ;-)

imj on 25/6/03 6:49PM [ Reply | Report ]

imj>So man, what's your problem then.

What horrible thing did Castle do to you?

I mean I can't honestly see much difference between RO 4.02 (which I am happy enough with) and RO 5.XX.

Where there *are* differences they are generally in RO 5.XX's favour (like allowing for ultra-large drives, using UDMA at many times the speed of the old IDE on the RPC). Support for the xScale means that we're not likely to be left high and dry without a processor anytime soon. Also with a degree of hardware abstraction it means "retargeting" the platform for other uses becomes a lot easier (assuring the platforms future).

Now granted Castle needed to put a HAL in there to allow the Iyonix to run (while I am not a major fan of HAL's - they can sometimes hit performance - Castle seem to have got the balance right). The machine I saw in Guildford last year seemed quite fast.

You ended by saying "you can't say an Iyonix is a good replacement for a RISC PC", that I would dispute - but I'd also add "the Iyonix is the *only* replacement for the RISC PC" - and that my friend is the reality of the situation.

-- Annraoi McShane,

AMS on 25/6/03 7:22PM [ Reply | Report ]

IMJ: I asked you weeks ago to justify this nonsense you keep spouting about RO5. You're a troll!

N.

NeilWB on 25/6/03 9:04PM [ Reply | Report ]

MD cant use the discontinuation of the SA110 as an excuse as there are still thousands in the distribution channel, and you will be able to source small quantities of the chip from suppliers for many years to come.

Luckily there are very few of the last SA110 revision that wont overclock to 300MHz. so even this shouldn't be a problem for the tiny number MD will actually require.

---druck [link]

druck on 26/6/03 1:19AM [ Reply | Report ]

"Where are Gilbert & Sullivan when you need them ?????"

Did you have a particular bit of G & S in mind?

e.g. "Misery me, lackey-day dee" or whatever.

Martyn

Martyn Fox on 26/6/03 7:24PM [ Reply | Report ]

Any song and dance would have done, even the Pirates of Penzance

-- Annraoi McShane,

PS: I really think I am beginning to lose it (one hamburger in two years and now THIS....)

AMS on 26/6/03 7:38PM [ Reply | Report ]

I realised on hearing Jack Lillingston and David Atkins talk at Wakefield that their companies have totally different design philosophies.

MD appears to have been developing its own proprietary hardware for RISC OS (though as part of its broader "Spectrum" project). Castle, OTOH, has been using "industry standard" hardware (notably the graphics card) and tweaking the software, by adding on the HAL, to run with it.

Martyn

Martyn Fox on 28/6/03 4:59PM [ Reply | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Create a new account
Forgot your password?

Search this website