This article is more than 1 year old
Pentium 4 high risk strategy for Intel
The thrashing of the Pentium 4?
Many of our lovely readers have drawn our attention to this piece at emulators.com, a lengthy diatribe about the Pentium 4 written by one Darek Mihocka.
Essentially, it looks at the Pentium 4 and its architecture in quite some detail and you can read for yourself what Citizen Mihocka has to say, although it's possible to gauge the general tenor of the piece from its opening paragraph which reads: "According to Gateway's web site, the Pentium 4 is 'the most powerful processor available for your PC'. Unfortunately for most computer users, it's simply not true."
As we reported over the last two weeks, Intel will be pushing the Pentium 4 with all its might in the opening year of the third millennium, employing, let there be no doubt about it, a barrage of marketing techniques which it hopes will make the chip the Numero Uno x86 processor, irrespective of any blemishes it has.
We've already detailed some of these marketing strategies, which include quick drops in pricing, a swift ramp of higher clock speed P4s, kissin' and a cuddlin' its PC customers, and ensuring there's enough of them thar Rambus RIMMs at low enough prices to make systems affordable.
Is this a high risk strategy for Intel?
The Scary Sandmen are, believe you us, quite prepared to take risks, knowing that if the gamble comes off, the gravy train will keep on rolling, the boss-men and the shareholders will get their bonanzas, and the capital expenditure and R&D can keep on cranking.
These semiconductor folk, with few if any exceptions, whether they be from Intel or AMD, are aggressive, single-minded and driven individuals who don't have laid-back Redmond or Cupertino airs and graces.
We have lost count of the number of projects Intel, over its history, has introduced and quietly canned, often at fantastic expense. Some, like Mr. i960, and more recently Master 710, never really made it. Others, like pulling the rug from under 3Com's feet by capturing the network interface card (NIC) market by stealth, have been super soaraway successes making a lorra lolly for Santa Clara. The failures are airbrushed from history or consigned with their owners to Intel cubicles in its Griffin Island, Banff or Irkutsk outposts.
A pullet in the pen is worth a hundred in the fen
Citizen Mihocka writes, quite rightly, that the features of the Pentium 4 platform require software to be optimised for it to sing like a thrush rather than quack like a duck. He also makes the perfectly valid point that no amount of rewriting will make most of the software the world+dog currently uses run faster on the P4 than it currently runs on the Pentium III or on the AMD Athlon.
Many software compilers such as Microsoft's Visual C++ 6.0, he adds, still aren't optimised even for the Pentium III platform, so god alone knows when compilers will hook into Intel's Screaming Sindy II and take advantage of the NetBust architecture.
He concludes that what he describes as "marketing scumbags" have pushed out innovators at Intel, and urges the world+dog to boycott Intel.
We know some of these "marketing scumbags" at Intel and we think that they're actually very bright people who (a) managed to bamboozle the entire universe with the Intel Inside campaign, (b) have successfully launched a number of PR initiatives judging from much of the press coverage Chipzilla gets, (c) know how to do the same with the financial investor suits in Wall Street and (d) have even, at times, argued against wilder counsels at Chipzilla who wanted to sue the hell out of the ratpack, The Register and anyone who even dares to suggest Intel had screwed up.
Worse than this, at least from Citizen Mihocka's point of view, Intel's marketing engine does stand a chance of pulling the wildly weird Pentium 4 rabbit out of its conjuring hat in 2001, despite his perfectly valid arguments, mainly because it is prepared to take risks and put its marketing moolah where its mouth is.
Like we said in our Buyer's Guide the other day, a 1GHz Athlon or 1GHz Pentium III with oodles and oodles of very cheap SDRAM will be a very attractive proposition for some months. But when the Pentium III becomes a 'disappeared chip', what then? Distributors, dealers and PC manufacturers need to stick something into the machines wot they want to sell.
Opportunity is whoredom's bawd
And so to AMD. In the last few days, august organs such as the Wall Street Journal and others have started writing about how AMD, which managed to gnaw a big chunk out of Intel's desktop share over the last year, now has to complete the transmogrification by wooing the corporate marketplace.
Lest anyone forget, AMD has attempted to do so before, but was often hampered by the fact Hipzilla was playing ketchup with Chipzilla. The corporate marketplace required stability and until the likes of Atiq Raza and others injected an Intel-style atmosphere into AMD three years ago, trips to Hawaii and fantastic press parties (believe us, believe us), were all part of Jerry Sanders' thang.
There's no saying that AMD cannot or will not make inroads into the corporate marketplace, but this certainly won't happen overnight. AMD will have to prove itself at every single step and the buyers at the big corporations will be watching the numbers on Hipzilla's balance sheet as well as its stock price when they make their big decisions.
The server market could provide the thin end of the wedge for Hipzilla if it manages to get Sludgehammer out there with, for example, the backward compatible Win64 running on top, before Intel's Foster platform starts kicking butt.
Cache for questions
Citizen Mihocka makes some telling points in his piece on the Pentium 4, commenting, in particular, on the cache compromise Intel made before the release of the processor.
Readers may recall that when we first published specs of the Willamette some two years back, provided helpfully by an insider at Intel Israel, the idea was that the caches would be wonderfully huge.
We believe along with Citizen Mihocka that Intel was forced by the rise-and-rise of AMD technology, to make these compromises in order to get the processor out of the door. But we also have reason to believe that these innovations have not been abandoned on some dusty Oregon shelf, and that Intel will spend its marketing money to establish the Pentium 4 before bringing back the things that perhaps should have been there in the first place.
And our readers write...
Citizen Mihocka has certainly stirred up the juices out in Register Readerland. We've had quite a response so far to his sayings, and will wrap them up in a little parcel tomorrow. Read the original rant and let us know what you think. So far, the emails we've had offer quite some few corrections to the views of said Citizen. ®