Electronics Weekly Electronics Design & Components Tech News
Intel’s Medfield chip-set aimed at getting the company into smartphones has been dissed by ARM CEO Warren East.
“It’s inevitable Intel will get a few smartphone design wins,” added East.
Getting up close and personal, East added: “They have taken some designs that were never meant for mobile phones and they’ve literally wrenched those designs and put them into a power-performance space which is roughly good enough for mobile phones.”
Intel reckons Medfield will get it into smartphones and tablets.
Qualcomm says its ARM-based Snapdragon S4 has won 70 non-handset design wins at 20 manufacturers.
Battle has been joined.
I can’t remember the picoturbo saga, but would guess the ARM7 patents fall off the lorry in 5 or 6 years time.
It might be hard to kill off the ARM7 ecosystem, which is presumably what ARM is doing with the cortex range? To get another set of patent ammunitions.
Majority of embedded applications would be fine with an ARM7 (or even earlier) architecture. Nevr worry about Java bytecodes or thumb – just straight 32bit machine-coce.
It is inevitable the trusty ARM ISA will go the way of the 8051.
It’s more complicated still – if US patents were filed before 1995 and have gone through divisional or continuation applications there could still be unpublished, unexamined submarine patents waiting to torpedo anyone who tries to clone ARM 7.
It is unlikely, and also risky that they would be enforceable, given what happened in the Lemelson vs Symbol Technologies & Cognex Corp case, especially as ARM isn’t a US company and probably wouldn’t have much political support for any legal case.
That’s quite a good point, Stooriefit, but the other reason for not using all the interview is because I can’t. The rules for lifting stuff from other sites are that you can take a selective snippet or two but that’s max. You have to limit it, or there’s trouble. Besides which, the emollient add-ons were obvious – obviously Intel is a competitor, obviously it has more to offer than power-efficiency. Being obvious they didn’t add anything – except emollience
Being mischievous, I could interpret “the story lay in the criticism of Intel’s part” as a journalist’s excuse to selectively quote and “and the other bits were emollient add-ons” as just an emollient add-on.
I wonder what Goffers thinks of the full quotes?
Not quite, the US rules are different for patents filed before 1995.
Expiration date is 17 years from the date of issue or 20 years from the date of application of the first patent in the family whichever comes first.
Thumb-2 came with ARM1156 core and was announced in 2003.
To me thumb opcode shows that the old battle RISC vs CISC has clearly been won by CISC, since a dual instruction 32bits opcode is inherently a complex opcode and not a RISC one despite what ARM acronym stands for.
With today’s bottleneck often being memory bandwidth, compressing as much information as possible in the smallest area seems the way to go.
Thank you, Mike.
Thanks, Mark
The Thumb compressed op-code extensions which everyone uses were added much later though
The ARM v7 ISA only came along with the Cortex series of processors. The ARM7 cores ran ARM v4T ISA.
20 years from being granted, Lefty, as I’m sure you know, and the launch date of ARM 7 was 1994.
How long before the ARM patents expire and the ARM7 ISA becomes a clone target? (just as how the 8051 ended up)?
Because, Mark, the story lay in the criticism of Intel’s part and the other bits were emollient add-ons.
Why have you only used partial quotations from the original Reuters interview?
e.g. “It’s inevitable Intel will get a few smartphone design wins” creates a slightly different impression to the original “It’s inevitable Intel will get a few smartphone design wins — we regard Intel as a serious competitor”
or “Are they ever going to be the leaders in power efficiency? No, of course not.” reads a litte differently to the original “Are they ever going to be the leaders in power efficiency? No, of course not. But they have a lot more to offer.”
Best comment for Intel from ARM
He`s basically said Intels first rough put-together designs has already caught up with our years of experience.
I bet we wont be hearing any other comments from the ARM CEO soon about Intel.
The piece with the picture of the Grim Reaper, Lefty, on the assumption that Intel can only get x86 everywhere if they kill ARM and can only kill ARM by buying it. Intel doesn’t want to make ARM processors – they tried that.
why would Intel buy ARM ?
Intel can simply licence the design/arch and extend from there same as everybody else does ?
ARM can make a few pennies on an IP block, while Intel make an order of magnitude more coz they actually sell ‘puter chips.
Intel should concentrate on their expertise in building chips.
what puzzle piece am i missing here?
You’re absolutely right IMHO, Geoff, and I notice that the business model line was the one taken by Hermann Hauser last year when he told the Wall Street Journal that ARM could take on Intel in the computer business. However, as you say, Warren East has always put the view that it’s ARM’s power-efficiency that will win the day. To my mind this is an advantage that will wither. Intel has to catch up one day unless its technologists are staggeringly delinquent. And when Intel catches up on power-efficiency then, as you say, it becomes a battle of business models (and the power of Intel MDF). So marketing will be very important and, in this regard, ARM are awful.
I think ARM places a little too much emphasis on its technology as its differentiator. When commenting, they almost always fall back to this knee jerk response. Intel has the capability to out-process ARM and its ecosystem at the moment, it now needs to find a way to take its sw ecosystem legacy forward on a processor architecture that is more power economical. they have the cash to develop such a solution. So i am not counting them out yet.
Perversely, the real strength of ARM is its ecosystem and all the business aspects of that ecosystem in terms of vested interest by other players, who combined, can hold off Intel.
I have always said the intel ARM battle is one of business models. ARM is underplaying its hand in this game in its marketing by a long way in my opinion. ARM and its ecosystem will suffer if they don’t pull together to face the very real Intel threat to their mobile processor business.
Yes, Goffers, you’re not the first to comment on a new and unfortunate note of cockiness emanating from the ARM guys. Like you, I just hope it doesn’t affect their business.
I’m an admirer of ARM and its management, but I must say that I’m not very happy with the rather arrogant tone of some of Warren East’s comments in the last year or so.They say that pride comes before a fall; I hope that doesn’t happen to ARM.
Intel has declared things throughout its history. Some happened (e.g. plesiosynchronous and packet voice networks, graphics accelerators, WiFI, USB, ADSL), some didn’t (e.g. WiMax, universal wireless broadband).
But they are still the #1 semiconductor company with a process 2.5 years ahead of anyone else and growing. Of course an ARM processor on an Intel process would be the best solution but that isn’t going to happen anytime soon.
And yes, a bidding war for ARM would be interesting indeed.
You are right David.
There are a number of monster players with a vested interest in making sure Intel don’t control ARM, and amongst them Apple have the biggest cash pile.
They are currently holding about $10B more cash than Intel… ($15.20B vs $25.95B according to Yahoo! Finance).
It is the prospect of the bidding war which could ensue that is keeping ARM’s market cap up at $12B on revenue of $750M.
Wouldn’t be surprised if ARM’s valuation didn’t approximately track a proportion of the size of Apple’s cash pile, but I’ve not checked the history.
Well, Mike,15 years after declaring its ambition to be the pre-eminent supplier of building blocks to the Internet economy, it isn’t. So technology is all very well but applying it to customers’ needs can be a different matter. While buying a company for $12 billion (ARM’s market cap) which has revenues of $750m might annoy Intel’s shareholders.
Hmm. Possibly not the best of moves. Never, ever under-estimate Intel’s technology.
Or the fact they could buy every share in your company out of petty cash 🙂