I've been going through the various reviews, trying to develop a basis for whatever analysis I put forth on the P4 vs. AthlonXP match up. So far, it's shaping up to show only one thing with clarity, the 2.2Ghz NW P4 and the AthlonXP 2000+ are pretty much dead even, neck and neck. In fact, I am surprised so far just how evenly split the "benchmark wins" are between the two chips from review to review. There really isn't an edge in either direction that I can see so far. __________________ TMF Money Advisor
However, that's not the point of this post. I just had an epiphany on Intel and the P4 that I wanted to share while it's still fresh in my mind.
Intel really took a huge gamble on the P4 core design, which doesn't seem like Intel, but here we are. The gamble was that raw Mhz was the end all and be all. If they could dominate the competition in this one area, all else wouldn't much matter. They wanted the Mhz lead so badly that they were willing to make some pretty serious trade offs to get it. The P4 is now running at 2.2Ghz, but at what price?
First off, Intel sacrificed clock for clock performance. We all know this and it's why we are in the situation as it stands today, where Intel has a 533Mhz lead, but still can't reclaim the performance crown.
Has this paid off for Intel? Well, initially it looked good. People bought into the Mhz Myth, Intel was able to maintain a healthy price premium at Mhz ranges AMD couldn't touch. However, things are changing. The Mhz myth is being broken down with each passing day, in spite of initial moans and groans, AMDs speed rating seems to be working like a charm.
There is an additional cost to Intel in the Mhz smoke and mirrors, people are starting to understand that not only does the P4's 2.2Ghz mean a performance advantage over the 1.67Ghz AthlonXP, but that the P4 1.4 Ghz they shelled big money out for at launch didn't really offer them much extra performance over a 1Ghz PIII. There is the risk that as consumers understand the trade off Intel made to offer the "illusion" of greater performance, there could be a backlash.
Performance wasn't the only thing that Intel sacrificed. The die size of the P4 was huge on the .18 process. Even though the Northwood looks more reasonable, it is still LARGER than the AthlonXP, even with the process size advantage. That advantage disappears soon, when the .13 Northwood will be almost TWICE the size as the .13 Athlon core. That is a price that Intel paid for its Mhz first strategy that should not be ignored.
Larger die size equals greater cost and potentially poorer yields, as there are more transistors that can be faulty.
Intel traded two large disadvantages, performance and die size, in a gamble aimed at returning to the dominant position in CPU design and performance. How has that worked out for them? Even at the brief window in time where Intel is shipping a .13 product and AMD is not, AMD still has managed to equal Intel in performance, while maintaining a smaller die size, even with the process size disadvantage!
To me, that proves that Intel's gamble has been a failure. It's clear why they opted for the larger L2 cache, in spite of the extra die space. If they hadn't, they wouldn't even have been able to match AMD on performance, while losing out in the other two major areas of chip design. AMD's .13 Athlons are a quarter away, and we should be seeing them right on time to match Intel's next speed grades. AMD should easily be able to maintain performance parity through out 2002 and clawhammer should bring CPU performance and design to a level that Intel seems to have little chance of matching.
Got money questions? Your answers are just a phone call away! TMF Money Advisor puts you in touch with an objective Financial Planner whenever you need it.
I've been going through the various reviews, trying to develop a basis for whatever analysis I put forth on the P4 vs. AthlonXP match up. So far, it's shaping up to show only one thing with clarity, the 2.2Ghz NW P4 and the AthlonXP 2000+ are pretty much dead even, neck and neck. In fact, I am surprised so far just how evenly split the "benchmark wins" are between the two chips from review to review. There really isn't an edge in either direction that I can see so far.
TMF Money Advisor