Analysis: Despite Intel’s openness, Apple keeps us guessing
As positive as the switch to Intel-built processors has been for Apple, one aspect of that partnership must drive Steve Jobs’ control-freak nature absolutely bonkers: secrecy. Or, more to the point, the lack of it, when it comes to future product announcements.
Prior to Apple’s 2005 decision to use Intel chips in its hardware, the Mac’s processors were provided by IBM and Freescale (née Motorola), and Apple could tightly control what little—if any—information it provided about its products prior to their introduction. Neither chipmaker released roadmaps of their product lines that were anywhere near as detailed as Intel’s.
Intel, on the other hand, keeps its partners—and, by extension, the general public—well informed about its product plans years in advance. Although this openness insulates the entire PC ecology from expensive surprises, it can also inflate customer expectations while delaying purchases. After all, if you know that a much more powerful chip is going to be released Real Soon Now, why write a check today?
Tim Deal, senior analyst at market-research firm Pike & Fischer, acknowledges that Apple formerly was able to keep its product plans “a closely-guarded secret” and that the company “tightly managed and exploited this secrecy to create anticipation (and drama) around its new product releases.” But he believes that the new Intel-enforced openness is a good thing.
“[Apple’s] reliance on the not-so-secret Intel processor development cycle allows the market to get a glimpse of future processor incarnations for the Mac,” Deal said. “This does not inhibit the company’s ability to generate excitement about future product releases nor does it preclude Apple’s customary mystique and showmanship surrounding a given product release.”
What’s more, when Apple was tied to PowerPC processors with lower clock speeds than competing Intel chips, it struggled to explain away what it then called the “Megahertz Myth.” As Deal explains, “Consumers are now able to compare Mac computers alongside of PCs with greater effectiveness.”
What’s in a name
That’s true, of course, only if customers can decode the sometimes arcane world of Intel roadmaps, codenames, product lines, and platforms. Recently, for example, came the announcement that a six-core Xeon processor codenamed Dunnington would be released in the second half of this year. The mere codename of this chip speaks volumes about the difficulty of deciphering Intel’s roadmaps—Dunnington’s original codename was Aliceton, but it was renamed after the development of another Dunnington processor was halted when a third processor-development effort—codenamed Whitefield, upon which the first Dunnington was based—was discontinued.
Confused? You’re not alone. But worry not, Dunnington lore won’t be on the test—that six-core beast is unlikely to ever make it into any Mac Pro. Mac Pros use Intel’s 5000-series Xeons, which fit into their dual-processor platform codenamed Stoakley. Dunnington is a 7000-series Xeon, designed for the quad-processor, server-level platform codenamed Caneland.
You’re also to be forgiven if you were never informed that Intel and Apple have two different definitions for the term “desktop processor.” When Intel introduced its Core 2 Duo processor line, the first desktop processor out of the gate was codenamed Conroe. This chip, however, never made it an Apple “desktop” computer. The iMac was instead built around what Intel referred to as a “mobile processor,” codenamed Merom.
Merom’s replacement, which shipped in November of last year, is codenamed Penryn, and it comes in many flavors with more to come. For example, Intel confirmed last month that it will release a quad-core Penryn later this year. Although this chip is identified by Intel as a mobile processor, its power appetite (defined as its TDP, for Thermal Design Power) is expected to be a hefty 45 watts. In comparison, the TDP of the new Penryn-equipped MacBook Pro is 35W—and if you’ve had one on your lap for any amount of time, you know it’s already pushing the thermal envelope. The increased TDP would also, of course, more-quickly drain battery power.
That said, there are plenty of 17-inch MacBook Pro-using video editors who’d kill for those two extra cores when encoding on location, plugged into AC or not. The 17-incher also has more room for cooling technologies and batteries, should Apple care to make use of those extra cubic inches.
The core of the matter
A quad-core iMac isn’t out the question, either—but this talk of multiple cores begs another question. Namely, are extra cores worth the extra money, heat, and power? The answer, of course, is “It depends.” Each core needs to be fed enough data and instructions to keep it humming along efficiently, and many if not most applications these days aren’t smart enough to take full advantage of multiple cores.
Intel is working on a pair of technologies to help the transition into the multi-core world: SPT (Speculative Parallel Threading), which will work to exploit all cores in real time, and TM (Transactional Memory), which will ensure that the work threading its way through one core doesn’t muck up the work done by another core. However, the real solution to multi-core management will be highly optimized software, and that’ll take some time.
What’s more, multi-core machines—think eight-core Mac Pros, for example—work best when their storage systems can provide torrents of data to their processors. A Mac Pro can be tricked out with a fast internal RAID array. Not so a single-drive quad-core iMac.