If you take a look over here, you’ll see a chart showing purported Sempron PR ratings, frequencies, and a few other sundry details.
When you first look at them, what catches your eyes is that they have higher PR ratings than normal Athlon XPs/Athlon64s, despite running at lower frequencies and/or having less cache than Bartons/XPs/64s.
My initial reaction is “How terrible. AMD is perverting the PR system.”
Then my brain started working, and I realized what AMD was doing.
Semprons are being matched up against Celerons. When it comes to perverting the meaning of gigahertz as an objective measurement of performance (granted, a false notion, but a common one), Celerons have been to Athlons what necrophilia is to reading Playboy.
The situation has improved somewhat with Prescott Celerons, but Intel has perverted the concept of gigahertz as some objective measurment of overall performance far more with Celerons than AMD ever did with PR, simply by keeping quiet about it.
Ironically, it did this so effectively that Intel hurt itself by it. Plenty of people thought and bought on the basis that there was no real difference between a Celeron and a PIV besides price.
So what AMD is doing is not creating new perversions, but matching Intel’s old ones in any head-to-head matchup. This makes sense, and in the battles of the Best Buys and CompUSAs, AMD is giving customers a level playing field with which to compare Semprons with Celerons.
Of course, by doing that, it gets rid of any direct correlation between Semprons and Athlon XPs/64s, but that’s not the important comparison AMD needs to make.
When Semprons do come out, and the numbers are all official, we’ll put together a chart comparing AMD to AMD.
For now, just know that what AMD is doing is overall a good thing for its intended audience, and if you don’t like it, blame the original perverter: Intel, not the matcher.