Intel the Silverback......stumbling about in the mist???
Can't say that I didn't warn you:
Message 14147871
Message 14148060
Message 14150735
Here's a few random thoughts on the Intel saga:
1. The holiday season, AMD will have 1.2-1.4 Ghz. Mustangs competing against 1 Ghz. PIIIs from Intel. The Mustangs may possibly also have the benefit of utilizing DDR. Even if DDR doesn't get rolled out in time, the performance advantage is in AMD's favor.
2. The release of the Mustangs will allow AMD to slash prices on its 1-1.1 Ghz. Athlons, which should make them very popular in the retail market, and possibly make some all-Intel corporate buyers think twice. Intel can't follow suit, since the 1 Ghz. PIII will remain their high-end product, and since Wall Street's already paying close attention to Intel's profits.
3. The PIV, when it comes out, will have twice the die size of an equally clocked PIII, which will significantly increase both its cost and the amount of fab capacity its production requires, the latter of which will further add to costs as Intel will remain capacity constrained until the middle of next year, when they switch to .13 micron (AMD's expected to switch around the same time). Meanwhile, the larger heat sink and fan, and the $70 rebate on PC purchases also should contribute to a steep ASP for the PIV when it first comes out, and should relegate it to the status of a high-end niche product until the switch to .13 micron.
4. Intel's hoping to justify the higher costs that'll come with the PIV by means of offering superior performance for resource-intensive multimedia apps. Even if it's assumed that the PIV will ofer superior performance for such apps, it should be noted that the collapsing ASPs of the past two years have shown that the majority of PC buyers, both in the retail and business sectors, are only willing to pay for superior performance at the right price. The slow adoption of RDRAM is a testament to this fact.
Also, with regards to multimedia apps, streaming video and DVDs aren't widely used in business situations. Voice recognition is occasionally used, but is far from attaining mainstream status. Meanwhile, the extent to which price premiums will be accepted within the consumer market's also debatable.
A few days ago, after running a DVD, playing a couple of graphics-intensive games, doing some Photoshop tests, and watching 30 minutes worth of streaming video at 300 kbps, all on my 550 Mhz. Athlon, I'm having trouble buying the arguments that even after more than doubling the clock speed and substantially increasing the amount of memory bandwidth available, you'll need specialized instructions to handle such applications. While I occasionally saw some choppiness in the video frames that were displayed, this seemed to be more of an issue related to the video card rather than the CPU, the latter of which didn't appear to be heavily taxed at those points. In fact, the biggest CPU-related problem wasn't the handling of high-quality graphics or video, but the loading of the software that handles such apps (i.e. Photoshop, WMP, Real Player) in a seamless, instantaneous manner. However, what's needed here's standard, run-of-the-mill processing power, something that AMD appears set to offer at a much lower cost than Intel.
5. Intel's hoping that the use of RDRAM in conjunction with the PIV, combined with the huge front-side bus that's capable of fullying exploiting RDRAM, will allow it to attain a performance edge. I'm inclined to agree with those who think that long-term, DDR won't be able to scale as well as RDRAM, and that AMD will eventually have to create motherboard chipsets for its processors that support RDRAM. However, 266 Mhz. DDR should be more than adequate for the time being, not to mention much cheaper; and considering that AMD's actively trying to hire engineers with experience in RDRAM, it seems that they are planning long-term.
6. Intel still dominates the laptop CPU market, but the Mobile Athlon comes out early next year; and if the price/performance advantage AMD has is comparable to the kind it has in the PC market, there could be some problems for Intel. Granted, AMD hasn't released this chip yet, and they could screw up royally in this endeavor of theirs.
7. Intel's counting on the Itanium to propel its CPU-related growth going forward. The Itanium has been created from scratch to handle 64-bit applications. AMD, meanwhile, has opted to create a backwards-compatible 64-bit architecture. The problem with Intel's approach is that if Microsoft hasn't released a 64-bit version of Windows 2000 by the time the Itanium comes out, it'll most likely underperform the Sledgehammer on this platform, and will be primarily relegated to Unix systems. Meanwhile, here's a couple of good articles on some of the technical shortcomings of the Itanium:
fool.com
fool.com
Now, ignoring these articles, let's assume for a moment that the Itanium still does outperform the Sledgehammer. Since it's a built-from-scratch 64-bit design, it'll still most likely cost a lot more. This isn't a problem in the high-end server market, where companies are more than willing to pay a price premium. However, this market's dominated by Sun and IBM, both of which use their own proprietary CPUs; and I doubt that the Itanium will have a significant amount of success cracking this field. So, like the Xeon, for the most part, it'll have to compete in the low-end and mid-tier server markets, where price/performance does matter, and where it'll face a solid competitor. Also, the potentially lower cost of the Sledgehammer could make it a viable threat for the PIV in the high-end workstation market. In all honesty, I do have to admit that Intel's clout in the corporate arena could make it difficult for AMD to crack a large number of accounts
To sum all of this up, it seems to me that Intel's acting as if it's still 1995, back when it had a virtual monopoly on the PC and Wintel server microprocessor market, and the ever-increasing complexity of software was driving both business and retail buyers to pay top dollar for PCs without anywhere as much of a regard for costs. While the company keeps trying to look to the future by means of entering new markets, I'm not sure if it fully understands that the dynamics surrounding its older markets have changed dramatically. The history of the IT industry's filled with great companies that unraveled as a result of making a similar mistake.
Eric
PS - Just remember, Intel's still primarily a microprocessor vendor. All of the company's non-CPU-related businesses combined for an $800+ million loss last quarter.
PPS - Intel longs should take a look at some of the comments on the message board of this site:
faceintel.com
While a lot of the comments there are standard, populist, anti-corporate BS, some of the statements provided by individuals claiming to be current or former Intel employees are quite startling. |
