Drobe :: The archives
About Drobe | Contact | RSS | Twitter | Tech docs | Downloads | BBC Micro

Sir Robin leaves ARM

Published: 27th May 2006, 20:18:16 | Permalink | Printable

Meanwhile another ARM11 core licensed

Sir Rob SaxbyARM co-founder and chairman Sir Robin Saxby is stepping down this year from the chip design giant. He had also served as chief executive officer and president since joining the Acorn spin-off in February 1991.

The move somewhat marks the end of an era at ARM, which was created at the start of the 1990s when Acorn spun out its RISC processor design side - Acorn, Apple and VLSI Technology each initially held shares in the new company.

It was also the same company that triggered the break up of Acorn, which towards the end of the 1990s was sitting on a pile of ARM shares. Today, the processor design business employs 1,400 and enjoys an annual net income of £42m. Last year, 1.7 billion chips featuring ARM cores were manufactured, the architecture making up some 75% of all 32-bit embedded CPUs.

Sir Robin will leave the company in October.

He said: "Throughout my life I have been passionate about the importance of creating value for stakeholders from technology... I am delighted to be able to continue to act as an ambassador for ARM globally and hope to encourage young people to take up a role in engineering, science and technology, as well as broadening my own knowledge of other engineering disciplines."

• Earlier this month, chip design firm Sunplus licensed an ARM11 core to use in processors aimed at consumer goods, including PDAs and digital cameras. The ARM1176JZ-S has a clock speed between 330 and 550MHz, and has an optional floating point maths processor. ARM also launched its new Cortex-R4 processor core (not to be confused with Castle's early attempt at embedded engineering). The new design can manage clock speeds of up to 400MHz.


ARM website

Previous: New games database for emulator users
Next: Ovation Pro gets Dutch translation


Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

I'm probably going to be moderated down for this, but is anyone else thinking of the song in Monty Python and the Holy Grail?

"Brave, brave, brave, brave Sir Robin! When danger reared its ugly head he bravely turned his tail and fled..."

 is a RISC OS Userthesnark on 27/5/06 11:01PM
[ Reply | Permalink | Report ]

Actually all things considered Robin Saxby did rather well. The ARM did succeed rather spectacularly - and the fact the shares in ARM rose in value so high considering their rather modest income never ceased to amaze me - I can only put this down to the spin (and salesmanship and deal making of Saxby).

As to ARM being responsible for Acorn's breakup that's I suspect a tad unkind. Acorn broke up primarily because of it's *management* they never really valued properly what they had (nor their customer base) and simply chose the "path of least resistance" and closed down and cash-in rather than compete. If they had done more *sooner* then the sad withdrawl of Acorn might never have happened - to simply lumber the blame on ARM for this is to somewhat miss the point.

Are their lessons in this for ARM, yes. If ARM does *not* respond to Intel's low power x86 initatives then ARM (IMHO) will go the way of Acorn. ARM *need* to get higher performance chips out there now that perform no less well in power consumption performance than a low power consumption x86 if they don't then their are in peril. Time will tell.

 is a RISC OS UserAMS on 28/5/06 3:14PM
[ Reply | Permalink | Report ]

He's probably very glad in hindsight he went into ARM rather than staying at Acorn.....

 is a RISC OS Usermarkee174 on 28/5/06 3:55PM
[ Reply | Permalink | Report ]

I don't think he ever worked at Acorn in the first place.

 is a RISC OS UserCogs on 29/5/06 2:14AM
[ Reply | Permalink | Report ]

He was approached by a headhunter appointed by Acorn to join their joint venture with Apple and VLSI Technology. So he was with ARM from the start and never worked for Acorn.

 is a RISC OS UserJohnB on 29/5/06 7:47AM
[ Reply | Permalink | Report ]


Thanks for the clarification. I thought he was one of those who was with Acorn and left when ARM was setup.

 is a RISC OS Usermarkee174 on 29/5/06 8:37AM
[ Reply | Permalink | Report ]


"If ARM does not respond to Intel's low power x86 initatives then ARM (IMHO) will go the way of Acorn. ARM need to get higher performance chips out there now that perform no less well in power consumption performance than a low power consumption x86 if they don't then their are in peril."

Intel's x86 processors are years away from being comparable to ARM's in terms of price/performance/power consumption. A much more immediate concern in the embedded market is PA Semi [link] and its low-power POWER processors.

 is a RISC OS UserWill! on 29/5/06 11:12AM
[ Reply | Permalink | Report ]


"A much more immediate concern in the embedded market is PA Semi [link] and its low-power POWER processors."

"The first PWRficient processor, a dual-core device, operates at 2GHz with typical power dissipation in the range of 5 to 13 watts"

"Will" 5 to 13 watts sounds like a lot for a PDA or phone to me. With a 9v battery thats 555milliamps to 1.44 amps.

I'd love to know what sort of battery you are going to use.

High Ghz => high power => low battery life. That's why I cannot see ARM processors being made for high Ghz.

As there is not enough of a market for Desktop ARM's I think we will have to wait a LONG timer before an Iyonix is bettered for performance.

 is a RISC OS UserJwoody on 29/5/06 12:04PM
[ Reply | Permalink | Report ]


There are plenty of embedded processors other than those used in PDAs and phones, and that's where the competition with Freescale, AMD and (possibly) PA Semi is most intense.

 is a RISC OS UserWill! on 29/5/06 12:57PM
[ Reply | Permalink | Report ]

"ARM need to get higher performance chips out there...": isn't the fact that ARM don't fabricate their own chips relevant here? Aren't decisions about clock speed for a given ARM reference design taken largely by the fabricators?

 is a RISC OS Userbucksboy on 29/5/06 6:12PM
[ Reply | Permalink | Report ]


"Aren't decisions about clock speed for a given ARM reference design taken largely by the fabricators?"

Not sure, I don't think you can just take a design, choose a fab scale and think you can speed up the clock because everything is closer together

Increasing the clock speed potentially brings in timing problems so I would have thought each design would have a maximum clock speed.

I know IBM had to change designs to get their PowerPC chips to a higher Ghz it was not just a case of changing Fab. Although puting things closer together helps.

 is a RISC OS UserJwoody on 29/5/06 7:16PM
[ Reply | Permalink | Report ]

bucksboy>Yes ARM don't fabricate chips - but they *do* set the "road map" for the chips that are produced. They *do* determine the instruction set, pipeline characteristics and so forth. If you look at ARM's road map you'll find often they don't really try to push the performance envelope (though I'd be more critical of them in the past).

Only relatively recently has ARM announced that larger L2 caches would be allowed - something that I'd argue they should have done some time ago. The Cortex also has some features that may lend itself to higher performance - but bear in mind it's ARM that put down the blueprint *first* then the vendors implement the real chips.

Bear in mind as well that ARM has a good relationship with many vendors I am pretty sure that they *could* have sample hardware produced to illustrate a point (for example the 5 core NEC/ARM design). Yes the talents brought by the Silicon founderies is important - but many of these companies have extensive experience designing complex and fast processing and I/O hardware so why not take advantage of that ?

Ultimately ARM's survival in the face of the competition I expect Intel to mount will not permit them to sit on their laurels. At one end you'd have lower power consumption but now increasingly fast microcontrollers from outfits like MicroChip and at the other end fast yet reduced power consumption x86's from Intel. Simply hoping that ARM can find a "niche" that fits between those two IMHO isn't an option.

 is a RISC OS UserAMS on 29/5/06 7:26PM
[ Reply | Permalink | Report ]


"Aren't decisions about clock speed for a given ARM reference design taken largely by the fabricators?"

In fact I don't think the fabricators get much say in the mater. They can choose to invest lots of money in a production line that uses a certain fab size. i.e. the size which they can resolve things too, normaly measured in nm. They can then take say an ARM design and produce it. They then test the chips cranking up the mHz at some points the chips will stop working correctly they then have to decide where they set the bar in terms of Mhz/Ghz. So some chips will pass and some will fail, this effects the yield they are going to get.

Generally the finer you can resolve things to the faster you can in general run the chips. But to go to higher resolutions costs mega bucks you have to either revamp a production line or build a new factory. A given chip design may also have to be changed due to timing problems otherwise you don't see the gain in speed you hoped for.

The junctions in Transistor hold a certain charge, its a function of the size of the junction and the material i.e. Silicon. The faster the Mhz the faster the charge in the junctions is pumped in and out, this results in more current which means more power which means more heat. Reduce the size of the junction and the faster you can go. More power means shorted battery life.

The other point is as AMS allured to. The faster a chip goes the bigger the impact of a cache miss is. So there is no point going to ever faster Mhz unless you make caches larger and memory faster. In the PC world as processors passed the 1Ghz mark, front side buses ( The speed to access memory ) where also increased and L2 caches got bigger, other wise all the increase in processor speed would have just been going to waste waiting for memory.

 is a RISC OS UserJwoody on 29/5/06 9:10PM
[ Reply | Permalink | Report ]

So this ARM11 core by sunplus would that be a good chip for desktops??

 is a RISC OS Userhighlandcattle on 29/5/06 9:55PM
[ Reply | Permalink | Report ]

A quick look round the internet with Google seemed to indicate that Intel and AMD have 65nm fabrication plants with one reference to Intel going to 45nm in 2007. The lowest ARM I could find was 90nm with some being 130nm. There is also an interesting article by Saxby outling the problems manufactures have going to sub 100nm in 2003 [link]

 is a RISC OS UserJwoody on 29/5/06 10:02PM
[ Reply | Permalink | Report ]


"So this ARM11 core by sunplus would that be a good chip for desktops??"

I doubt it, it sounds like it is designed for low power consumption devices like phones, digital camera's.

 is a RISC OS UserJwoody on 29/5/06 10:10PM
[ Reply | Permalink | Report ]

Cortex™-A8 processor which will revolutionize consumer and low-power mobile devices, enabling the delivery of higher levels of entertainment and innovation to end users. Launched at the second annual ARM® Developers' Conference, in Santa Clara, California, the ARM Cortex-A8 processor delivers up to 2,000 DMIPS making it ideal for demanding consumer products running multi-channel video, audio, and gaming applications. For next-generation mobile devices, the ARM Cortex-A8 processor delivers industry-leading performance and power efficiency while using less than 300 mW in 65nm technology. For the first time, low-cost, high-volume products will have access to desktop levels of performance using the Cortex-A8 processor.

I belive in the news release in 1995 they reckoned 1Ghz with 16 or 32k of L1 and configureable L2

Someone may be able to tell us what 2000 DMIPS means as it means little or nothing to me (however if it can give desktop levels of performance then surely itt would be a good choice for a desktop)


 is a RISC OS Usermrmac on 30/5/06 4:20PM
[ Reply | Permalink | Report ]

From [link] The Cortex-A8 processor will run at more than 600MHz in low-power 65nm processes with the core using less than 4mm2 of silicon (excluding Neon, trace technology and L2 cache).

High-performance consumer designs will run the Cortex-A8 processor at up to 1GHz in high-performance 90nm and 65nm processes

I guess the question is how many of the licenses e.g. Freescale, Matsus***a, Samsung and Texas Instruments will have 65nm low power production facilites for Mobil phones and if any of them have high performance 65nm production facilities.

But overall it does sound hopeful, I guess they should start appearing soon.

 is a RISC OS UserJwoody on 30/5/06 4:47PM
[ Reply | Permalink | Report ]

Haha Matsus***a can't be spelled. Drobe is like a nazi controlled state!

 is a RISC OS Userhighlandcattle on 31/5/06 10:07PM
[ Reply | Permalink | Report ]

This: [link],39020330,39272166,00.htm is the sort of market ARM needs to break into.

 is a RISC OS UserWill! on 1/6/06 1:52PM
[ Reply | Permalink | Report ]

highlandcattle: I once got kicked out of an IRC channel due to some ridiculous over-zealot language bot, for using the word "sniggers". I doubt even Drobe is going to be silly enough to censor that, but I'll see in a minute...

 is a RISC OS UserSimonC on 1/6/06 2:32PM
[ Reply | Permalink | Report ]

Come on, chris change this so we could spell company names!!!

 is a RISC OS Userhighlandcattle on 1/6/06 4:56PM
[ Reply | Permalink | Report ]

highlandcattle wrote > "Haha Matsus***a can't be spelled."

But they're also known as Panasonic.... maybe that would help (additionally it's a tad more pronounceable).

 is a RISC OS UserAMS on 1/6/06 8:10PM
[ Reply | Permalink | Report ]

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

Search the archives

Today's featured article

  • Archive booklets review part one
    MessengerPro, VirtualRPC, simple networking and programming guides
     20 comments, latest by jlavallin on 19/12/05 3:28PM. Published: 12 Dec 2005

  • Random article

  • News update
    What's been going down
     Discuss this. Published: 25 Jun 2002

  • Useful links

    News and media:

    Top developers:
    RISCOS LtdRISC OS OpenMW SoftwareR-CompAdvantage SixVirtualAcorn

    CJE MicrosAPDLCastlea4X-AmpleLiquid SiliconWebmonster


    RISCOS.org.ukRISCOS.orgRISCOS.infoFilebaseChris Why's Acorn/RISC OS collectionNetSurf

    Non-RISC OS:
    The RegisterThe InquirerApple InsiderBBC NewsSky NewsGoogle Newsxkcddiodesign

    © 1999-2009 The Drobe Team. Some rights reserved, click here for more information
    Powered by MiniDrobeCMS, based on J4U | Statistics
    "The Drobe 'Best of 2005 awards' seems to have been infiltrated by a form of favouritism or censorship"
    Page generated in 0.2158 seconds.