All Articles Tagged processors

On other mobile platforms (hi Windows Mobile!) we often spend quite a bit of time comparing the processors of different models, seeing which one is faster and seeing what happens when you set the clock speed of a given phone to a higher number. It’s “fun,” see, because not only can clock speed be radically different from phone to phone, but so can performance even on devices with similar clock speeds.

The nice thing about the iPhone: not doing that. Well, until now. MacRumors reports that the 2nd gen iPod Touch, though it sports the same processor as its siblings, actually has a clock speed of 532MHz compared to the rest at 412MHz. The result is that certain apps like TouchSports Tennis run much better on the iPod Touch 2nd Gen. Which ultimately meant that the developer had to optimize their app for the iPhone, iPhone 3G, and iPod Touch 1st Gen. All three actually perform differently with the game, with the original iPod touch falling furthest behind.

It’s a bummer, because as apps become more powerful and get closer to testing the limits of the platform, developers will discover that the ’single target’ advantage of the iPhone/iPod Touch platform may go away. We’re not talking about having to code for as many devices as you do with Windows Mobile or BlackBerry, of course, but it’s still worth noting.

Why not clock up the iPhone 3G. In a couple of words: “battery life.” In a lot of words, well, battery life plus when you have WiFi, bluetooth, Quad-band edge and Tri-band WDCMA (that’s GSM 3G to you) all packed together, they have to be finely tuned. Just changing the clock speed could be enough to ruin a whole raft of things beyond battery life. With smartphones, we live in a world where the choice of paint can radically alter signal strength, so it’s no stretch to say the changes in heat, radiation, etc. associated with a faster clock speed could potentially cause problems. Plus, again, battery life people.

It probably wouldn’t completely screw with everything if Apple upped the clock speed — people change clock speeds all the time on WinMo — but it can happen and it’s a bigger deal as they get packed more tightly together. These radios are packed pretty tightly in the iPhone 3G.

We reported, and reported again, and pondered, and speculated some more about Apple buying Palo Alto Semiconductor (PA Semi), but now Steve Jobs himself has put the rumors to rest with a pretty clear statement as to his intentions:

PA Semi is going to do system-on-chips for iPhones and iPods

Of course, like any Lost season finale cliff hanger, Jobs’ answers lead only to more questions. When will Apple switch from the current iPhone and iPhone 3G Infineon S-Gold chips to PA Semi? By the third gen device? Fourth? And who will PA Semi — a design company, not a fab — be working with to produce these systems-on-a-chip? Will Infineon be cut out completely? Will Intel’s new Atom platform never stand a chance on the iPhone? (Just the iTablet?!)

We know OS X is amazingly portable, having already run on PowerPC, Intel, and ARM architectures, so transitioning shouldn’t be a problem for the hardware makers, but what about software developers? Will we be looking at Universal App Store apps on day?

(Universal Apps are the name for applications that are currently compiled to run on both PowerPC and Intel Macs — with increased effort on the developers side).

And what makes an in-house design so compelling it will overcome the lost economies-of-scale Apple enjoys by using more widely adopted architectures? (One of the reasons they switched to Intel on the Mac side).

Ars Technica offers some ideas (and further analysis):

Though there aren’t any more details available than the quote above, it’s plausible to infer that the SoCs in question will be ARM-based. Some of the key members of the PA Semi team were formerly part of Intel’s XScale group, which designed a line of low-power, ARM-based embedded processors before being sold off as part of a company-wide restructuring.

Guess we’ll have to just wait (for WWDC 2010?) and see. Curiouser, as they say, and curiouser.

Read

Following up on Casey’s story this morning about Apple buying chip designer PA Semi, Valleywag (yeah, I went there…) brings another possible angle to the “yeahbuwhy?” table:

[The PA Semi chip's suitability for the iPhone] may well have nothing to do with why Apple bought the company. PA Semi’s prize is its founder, Dan Dobberpuhl, a famed chip designer, and his 150-person staff. At less than $2 million per engineer, the price Apple paid is in the range Cisco pays to snap up talented engineers. With them working at Apple, Jobs can push established chipmakers to adopt its technical innovations and perhaps swap licenses for intellectual property. That’s far more likely than actually switching away from Intel chips for the Mac; Apple actually explored using PA Semi’s chips before choosing Intel. Even the iPhone, which would benefit more from PA Semi’s low-power chips, is an unlikely candidate for an all-new chip design. Why? Volume economics favor Intel and Samsung so strongly that it’s hard to imagine that a new microprocessor design from the PA Semi team could replace their wares. $278 million doesn’t buy Jobs a rival chip; it buys him a tool to chip away at his suppliers’ prices.

Of course, other angles remain actually using the chip design (though PA Semi does not manufacture their own chips, meaning someone with a fab, like Intel, would still be needed), licensing the technology/technologies to someone like Intel to produce proprietary chips to differentiate Apple offerings (and make life harder for Hackintosh’ers??), or just to beef up the patent portfolio and put a little fear into Intel to, as Valleywag put it, increase their bargaining position.

Personally, getting the engineers and licensing the tech makes the most sense to me at this point, but who knows if come Macworld 2009, El Jobso will pull the back off a 4G iPhone to reveal a brush-metal PA Semi chip glittering inside? What do you think?