Options
Where Go Intel?
Victor Kartunov of X-Bit Labs takes a look at today's desktop computer market and analyzes what Intel should do, especially considering the threat from AMD's new Athlon 64 CPU's and the recently discovered problems with the upcoming Prescott CPU.
In spite of the diversification, the microprocessor companies are fishing most in the waters of the “mainstream” processors market. Such products sell in biggest quantities and bring in most of the company’s profit. The basic marketing tool for beating the competitor has been the performance showings in widely-accepted benchmarks. With time, however, the marketing departments were becoming ever more cunning (well, the “opposing” testers got more artful, too), while the benchmarks started to yield strange and ambiguous results. At the current moment, we’ve come to a situation, which is utterly incomprehensible for a beginner. We cannot definitely say who’s the leader now. The question “What is faster?” doesn’t receive a direct answer from even an experienced user or a journalist – you have to ask a few qualifying questions like “in what?”, “what application?”, and “what are you actually going to do with your computer?” Any experienced tester can easily come up with a score of tests that show this processor to be faster.
In some complicated cases, the results of one and the same benchmark will depend on the scene you use. I could name the LightWare 7.0b suite as an example. It came with two scenes, and it was the choice of the scene that determined which processor will be faster in the benchmark Thus, benchmarks do not simplify the situation at all. Instead, we have a situation when tests results exposed in two pages require twenty pages of commenting upon. In other words, our usual speed measurements have practically lost their pristine meaning. This is actually no “conspiracy” on the part of the manufacturers – the architectures of the modern processors and platforms differ so much that direct comparison doesn’t tell anything substantial…
This is similar to a potential buyer looking at the speedometers of a Land Rover and a Ferrari. Of course, the Ferrari will have bigger numbers, but will it always be faster than the jeep? Crossing the country? On a dirt road?
That’s exactly what we have with the processors. It’s now not enough to point your finger and say this one is faster. You have to be detailed – “it’s faster in such and such applications and under such conditions.”
Benchmarks being no longer important, the marketing departments changed their foils. They fence now with notions of heat generation, availability/necessity of SSE2 and, recently, 64 bit.
Catch the full editorial (10 pages worth) over @ X-Bit Labs.com
In spite of the diversification, the microprocessor companies are fishing most in the waters of the “mainstream” processors market. Such products sell in biggest quantities and bring in most of the company’s profit. The basic marketing tool for beating the competitor has been the performance showings in widely-accepted benchmarks. With time, however, the marketing departments were becoming ever more cunning (well, the “opposing” testers got more artful, too), while the benchmarks started to yield strange and ambiguous results. At the current moment, we’ve come to a situation, which is utterly incomprehensible for a beginner. We cannot definitely say who’s the leader now. The question “What is faster?” doesn’t receive a direct answer from even an experienced user or a journalist – you have to ask a few qualifying questions like “in what?”, “what application?”, and “what are you actually going to do with your computer?” Any experienced tester can easily come up with a score of tests that show this processor to be faster.
In some complicated cases, the results of one and the same benchmark will depend on the scene you use. I could name the LightWare 7.0b suite as an example. It came with two scenes, and it was the choice of the scene that determined which processor will be faster in the benchmark Thus, benchmarks do not simplify the situation at all. Instead, we have a situation when tests results exposed in two pages require twenty pages of commenting upon. In other words, our usual speed measurements have practically lost their pristine meaning. This is actually no “conspiracy” on the part of the manufacturers – the architectures of the modern processors and platforms differ so much that direct comparison doesn’t tell anything substantial…
This is similar to a potential buyer looking at the speedometers of a Land Rover and a Ferrari. Of course, the Ferrari will have bigger numbers, but will it always be faster than the jeep? Crossing the country? On a dirt road?
That’s exactly what we have with the processors. It’s now not enough to point your finger and say this one is faster. You have to be detailed – “it’s faster in such and such applications and under such conditions.”
Benchmarks being no longer important, the marketing departments changed their foils. They fence now with notions of heat generation, availability/necessity of SSE2 and, recently, 64 bit.
Catch the full editorial (10 pages worth) over @ X-Bit Labs.com
0
Comments
" Intel Go Down The Hole!!!! "
Intel Go down the Hole!!!!!!!!
Agreed
I love that quote... I'd forgotten all about it. And it still cracks me up, too...
The computer industry is struggling.
AMD smacked a home-rum.
Intel cursed when the ball went sailing over their heads.
Intel's waiting to come up to bat with LGA775 and the Prescott.
AMD might have to worry about their bat breaking when the pitcher throws the DDR-II ball.
Intel might have to drop the pop-fly and let X86-64 land on their watch, in their goods.
An then there's the new form factor, peripheral busses, chipsets...
It's good to be a geek, guys.
Mmmmmm, Home-Rum.....