There’s an interview with an AMD exec over here.
There’s two items of note:
The Gravy Train Is Over
We’ve been saying this for a long time, but here it is from the horse’s mouth:
“In the end, Fred left us with the following: if you see single core performance improving at a rate of 40% per 12 – 18 months, it will now improve at about half that rate for the foreseeable future.”
I think even that will represent a considerable challenge to both AMD and Intel.
If you haven’t already done so, lower the expectations, folks.
In A Prison Cell
The subject of Cell came up, and while you’d hardly expect anyone from AMD (or Intel) to say, “It’s all over, we’re dead meat,” the comments made indicate that the silicon giants don’t really get Cell. It’s the old looking at the new through old eyes.
Basically, whether you hear it from AMD or Intel (and they’re pretty much saying the same thing), it’s clear they’re both still slaves to the old geeky concept of “more, more, more” and think Cell is competing against whatever multicore complexity they come up with, and this (almost) completely misses the point.
Cell is not a competing technology. It’s a disruptive technology meant to change the paradigm of what a typical computer is supposed to be. Eventually, we’ll have big Cells, little Cells, many Cells, one Cell, providing a flexible framework in which 1) enough computing power gets delivered for the purposes of the (increasingly specialized) device and 2) everything from DVD players to workstations have the same core computing standards.
The point is not “We and Intel can build faster workstations from our CPU designs than you can with Cell.” That may well be true, it will almost certainly be so in the beginning. The major point to Cell isn’t to build a better workstation; it is to wipe out most of the workstation market because most people don’t need one, and would much rather have something cheaper/smaller/easier/fast enough to use.
Let me put it this way: Do you really think your cellphone-computer in 2012 is really going to need four or eight shrunk full-purpose CPUs to work?
The real battle isn’t really about hardware specifications; it’s about what computers are going to be like in the future. Maybe the best analogy for this is: Will they stay like PC, or become more like iPods?
This battle won’t be a quick fight. Unless the Cell people really drop the ball, this war will last years and years. Remember that the Cell folks have nothing to lose. They have a solid foundation in the gaming station market, and if they don’t come up with a PC killer the first time around, well, there’s always versions 2.0 and 3.0 after that. The folks at Microsoft can certainly tell you that you don’t have to get it right the first time to eventually succeed.