by Anton Shilov
03/15/2004 | 12:55 PM
Reports over a number of web-sites suggest that Intel Corporation is very likely to employ a new strategy to distinguish its chips from each other. If the claims turn to be correct, the giant chipmaker will break traditions it has been loyal to for about thirty years.
Historically Intel and a number of other chipmakers used megahertz to describe their microprocessors’ relative performance and to distinguish between the chips in one product line. Such approach proved to be pretty logical and viable, but there has always been another one – marking processors with numbers that reflect performance and/or particular processors’ position in its family. Traditionally Intel Corporation strictly used chip frequencies to describe its chips, while its main rival Advanced Micro Devices marked its processors with so-called performance ratings and model numbers.
With the introduction of powerful microprocessors based on totally different micro-architectures in the recent years, it became pretty clear that clock-speeds do not reflect real-world performance of processors or computers on their based. Furthermore, even processors with the same micro-architecture in general may offer different performance because of different sizes of cache as well as various speeds of processor system bus. As a result, the second approach of marking chips used in PCs became more logical than ever.
The Inquirer, PC Watch and News.com web-sites reported last week that Intel was going to follow AMD and mark its upcoming processors with a kind of model numbers that would reflect product positioning and relative performance within a concrete family of chips.
It is suggested that Intel Pentium 4, Intel Pentium 4 Extreme Edition, Intel Celeron, Intel Pentium M and Intel Celeron M microprocessors would be marked according to the new marking guideline that Intel seemingly wanted to deploy.
Central processing units from one product family would form three different series, presumably 3xx, 5xx and 7xx, where higher first digit reflects product positioning, e.g., “good”, “better”, “best”, while other digit would show relative performance of the microprocessors against other products in the series. This kind of approach would resemble that AMD uses with its Opteron processors for servers, where the first digit of the model number reveals a type of server the chip is intended to be used in, whereas other digits reproduce relative performance . Advanced Micro Devices also marks its high-end Athlon 64 FX desktop processors using the same strategy, however it substitutes the first digit that reveals product positioning with “FX” letters.
There were a number of reports about Intel's intention to deploy model numbers in the last 2-3 years, but none of them materialised.
Intel’s spokesperson declined to comment on the information.