highlandcattle
18/6/06 9:51AM |
Maybe some company will end up buying the Xscale division and make better use of it then Intel did |
AMS (+3.0)
18/6/06 3:56PM |
Good news, I spotted comments on this elsewhere. Really this actually a better outcome that previous news on Drobe would have suggested.
Of course we seen this all before when Digital Equipment Corporation sold its production facility (making DEC Alpha and StrongARM processors) to Intel (as part of a court settlement Intel lost as it happened....). Intel then proceeded to replace (after several years) the StrongARM with xScale and the IOP range (of which Iyonix's 80321 is one).
This is merely history repeating itself - and if Intel's ARM Architectural license is up for grabs too then the purchaser may have *considerably more freedom* to expand and improve on the performance on the xScale chip range. Overall I'd be optimistic.
In the interim Intel will run the division as is until the find a buyer - contrary to the news piece above Intel are *trying to sell the division as a going concern* - so again nothing to panic about.
What has RISC OS'es owner Castle got to do with "selling off the operating system if the price is right" got to do with this story ? RO5 is well positioned to be adapted to support whatever new chips the new xScale owner decides to produce. In fact this news may make it more likely for faster newer chips to arise - ie., not the best time to jump ship I would have thought. |
Jwoody (+3.2)
18/6/06 7:46PM |
highlandCattle AMS
I think you deserve OBE's for your optimistic attitudes, which I have to say seems to be a RISC OS users trait
Me I am a pessimist, I have doubts Intel will find a buyer. If DEC and Intel cannot make a go of things why should anybody else |
OliverB
18/6/06 10:44PM |
About castle selling off RISC OS, who were they thinking of selling it to? RISCOS Ltd probably couldn't afford to buy it, so who else did they have in mind? |
sascott
19/6/06 9:26AM |
In reply to OliverB:
Microsoft? Apple?  |
MikeCarter
19/6/06 9:50AM |
In reply to sascott:
nooooooo
In reply to OliverB:
How much is RISC OS worth? |
AMS (+1.0)
19/6/06 1:52PM |
In reply to Jwoody:
Thanks for the offer of the OBE .... I am impressed that Drobe attracts such influential people  .
Take a gander at the machines that use Intel's PXA's and xScale chips and you'll find a vertiable "who's who". The division is making an adequate turnover (250Million pa is not to be sneezed at) and if Intel don't get too greedy I am pretty sure they'll find a buyer.
Intel having an architecture license is a biggy - unlike many other players they're not just restricted to ARM7TDMI or 9 or whatever - they can build any and can where they want add facilities and change the processor pipeline. Having that freedom to exploit ARM - and having the extensive order book that they have may be enough to garner a sale - I would suspect from one of the more communications oriented companies.
In reply to OliverB:
I don't believe that Castle would sell it. If they did the next owner may not be as relaxed about RISC OS Limited pottering around and selling the OS in the background. So in the interest of both parties it's probably better that Castle hang onto it. |
highlandcattle
19/6/06 1:59PM |
The reason Intel is selling thei Xscale division is not the lack of profit. But because they are starting their own line of low-powerconsuming processors based on their x86 line. It would be silly to support two competing products. Now wouldn't it |
Jwoody
19/6/06 3:10PM |
AMS
250 million turnover sounds quite good but how much of it is profit? Thats the 64 dollar question. Also does not look good if you invested 10billion to get said turnover. I don't think greed will come into it. If Intel want a lot for the business its because they spent 10billion and want as much back as is reasonably possible. Greed would be asking for greater than 10 billion for the business.
Personally I remain pesimistic |
AMS (+1.0)
19/6/06 8:37PM |
highlandcattle wrote:
"they [INTEL] are starting their own line of low-powerconsuming processors based on their x86 line"
Indeed and this potentially represents a threat to all ARM based systems. ARM simply have to up the performance (speed) of their processors if they are to stand any chance of holding their market share.
In reply to Jwoody:
If INTEL are intent on getting rid of a less profitable division then they will. You should factor in that a great many companies have products totally dependant on what are (in effect) INTEL "proprietary" variants of ARM - the companies that use them can't simply "second source" a PXA270 or whatever (as Intel don't allow second source). This means that any organisation that "snaps" up the xScale/Comms/ARM division from Intel will have a somewhat captive audience (the old addage about it being easier to keep an existing customer rather than get a new one applies). And then there's the intellectual property too.
I guess we'll just have to wait and see - particularly as Intel really hasn't officially commented yet. |
Jwoody
19/6/06 9:20PM |
AMS
"You should factor in that a great many companies have products totally dependant on what are (in effect) INTEL "proprietary" variants of ARM"
I accept your argument about a captive market. One could even argue that Intel need to find a buyer so as not to alienate all those companies. But at the end of the day a buyer has to found, willing to invest a lot of money and with the sums involved be in it for the long term. I remain pesimistic until somebody buys into this, at the moment it looks like shades of the British Motor Industry and things being sold off. |
highlandcattle
19/6/06 10:02PM |
Perhaps ARM themselves could buy it back. |
Jwoody
19/6/06 10:16PM |
highlancattle
"Perhaps ARM themselves could buy it back"
Highly unlikely I think ARM decided a long time ago not to go down the road of manufacturing chips, its far too an expensive game for them costs in the order of billions.
Take a look at [Link: www.intel.com]
You will see that it costs over 2 billion dollars to build a fab factory, makes a 250 million turnover pale in comparision |
Cogs
20/6/06 2:38AM |
I think this thread is yet another case of people not bothering to read things properly and believing what they want to believe. Firstly the $250 million is turnover, not profit - the XScale division is a loss-making business. Secondly the $10 billion Intel spent was for a whole bunch of companies, not just the XScale division. And lastly, nowhere does it say the XScale division actually fabricates the chips it designs. Can we have a discussion on here actually grounded in reality for a change?
I can't see why the XScale couldn't be profitable either. ARM Ltd has never shown any sign of struggling, even when the rest of the industry was in dire straits and it's not as if their licensing fees are exorbitant.
I would look at Intel's webs***e, but it appears to be down. |
mripley
20/6/06 8:23AM |
What Intel will have is two low power products. One is their own design and the other is licensed from ARM for which they have to pay royalties. It makes economic sense to support your own license free product and sell off the other. |
druck (+2.0)
21/6/06 11:11AM |
I agree with AMS that this will be a positive outcome for the ARM ecosystem. Intel have vastly undercaptialised on the potential of the StrongARM, letting it stew for years after the DEC engineers left in disgust, and spending more effort on the rebranding as XScale as on the development of the chip. The architecture licence would be a huge advantage to a major ARM supplier such as Samsung or (ex-Motorola) Freescale, and they would fully exploit such an asset accross a range of low power and high performance devices.
Intel will live to regret the decision, as there is no way on earth to make an x86 more efficent than an ARM with the same physical process, as the just front end needed to convert the worlds least efficent instruction set in to something modern processors can actually execute, is bigger than any ARM core. |
epdm3be
23/6/06 1:15PM |
Oh please.
Intel IS using the tricks it learned from dissecting ARM-technology into their own designs. Eventually low-power x86 will surpass ARM-technology with the advantage of a much higher clockrate (which doesn't necesarrily translate into higher preformance gain as witnessed with the SA vs Xscale)
Anyway I believe is ditching their ARM business partly for reasons written by mripley and secondly because they now know enough of ARM to make their own low-power technology to compete with ARM (which is not-to-forget not a native US-company).
In fact there's NOTHING stopping Intel for doing excactly the same thing (IP-retail) that ARM does. Only they have vastly more money at their disposal.
Picture this: Intel buils its own IP for a low-power suposedly-high-performance cpu, ARM sees this and discovers ideas stolen from them. They sue Intel... by the time the US-judge states his verdicts (in favour of ARM perhaps), ARM is bancrupt. Intel buys the entire ARM Ltd. and throw it all in the bin. No-more ARM. No-more foreign competition.
And another US-monopolist gets bigger.
Naaah... this is fishy! |
druck (+1.0)
23/6/06 1:46PM |
In reply to epdm3be:
there are no inherent tricks in ARM technology for Intel to learn from, the reason why they are small and power effiecient is that they are simple designs that do not need all of the massively power expensive tricks for performance that Intel has to cram in to its chips running the bloated and outdated x86 instruction set. Anything executing x86 instrctions will always take more silicon and use more power than the equivelent clean simple orthoganal ARM ISA - theres no way physical round that.
Intel want to produce lower power x86s so Bill Gates can have his latest whim of running full Windows on a device a bit bigger than todays PDAs, and smaller then his previous latest greatest vision of the tablet PC which sank without trace. These x86s will be more efficent than todays ones, but they will never better the ARM or threaten its core markets, which consist of billions of devices which will never need to run Windows, and hence aren't chained to a power sapping legacy instruction set. |
AMS
27/6/06 7:30PM |
Apparently Intel have just sold the division that makes PDA chips (presumably including xScale) to an outfit called Marvell (nothing to do with the Fantastic 4 I hope .
Details that exist are on http://www.theinquirer.net/?article=32670
|
AMS
27/6/06 7:37PM |
Further to above on the Marvell site (I quote)
"After the close of the transaction, Intel intends to continue manufacturing products currently sold by this business for handheld devices and embedded applications, and to manufacture products that are being designed into upcoming devices. This arrangement is expected to continue while Marvell arranges other manufacturing resources. Intel and Marvell do not anticipate disruptions in the supply of these products due to this planned sale. "
(reference to link is [Link: www.marvell.com]!468540434?releaseID=581)
The gubbins were sold to Marvell for some 600Million USD.
So as I originally said *nothing to worry about* |
druck
29/6/06 9:23AM |
This is pretty much what we've been saying [Link: www.theregister.co.uk] |
AMS
29/6/06 7:35PM |
In reply to Druck:
Spot on (and thanks for the link).
Yes our impressions of the situation have been vindicated.
Real pity that the story was portrayed here twice by Drobe in a rather apocalyptic manner - just gets people worried when the facts actually don't justify such "lurid" commentary.
Still it probably makes for better headlines than "Nothing to see here, move along " 
|
| |
Intel has 