Thinking Outside The Box . . .

Many people justify AMD’s existence by saying, “If they weren’t around, Intel would rack up the prices and innovation would slow to a crawl.”

I came across this forum post, which gives such a different perspective to the conventional wisdom that I thought you ought to see it:

. . . . Some insist that if Intel becomes a sole monopoly . . ., the worldwide chip industry will stall to a snail’s pace.

I agree. It will. And it’s long, long overdue as well. It’s been moving far TOO FAST for more than a decade.

Software is a decade behind hardware. If Intel dragged its feet for a decade, it could REDUCE prices by not spending so damn much on new fabs every day and it would take 10 years for software to catch up.

Shrinks are rapidly running out of space and a slowdown would be welcome. Thus, AMD’s demise is not all bad.

Obviously, those reading this do not comprise the most receptive audience for this message. 🙂

However, it strikes me as containing more than a grain of truth about the future, regardless of whether that future includes AMD or not.

While it’s unlikely Intel will stop shrinking processes for at least another couple generations, it won’t be to increase performance, but to get more chips from a die.

Again, while we’ll see a few more CPU generations with improvements no matter what happens to AMD, Intel projects like Silverthorne indicate a change of direction away from bigger and faster to smaller and good enough.

Sorry, but the current “be fruitful and multiply” approach has inherent limits, which will be reached sooner rather than later. We’re not going to see Dell selling 16-cores with 16 video cards attached five years from now. That would just fuel console sales.

I think there’s something to be said for software catching up to hardware. Up to now, the software people have pretty much gotten a free ride from faster hardware, but what are the most common complaints about new versions of software these days? It’s about bloat; it’s about being slow.

At some point in the fairly near future (say three-four years from now), I think the burden of speed will pass from the hardware to the software people. It will no longer be a rising tide lifting all boats. Software writers are going to have to learn how to use multiple cores and write tighter, faster, smarter code if they want to keep adding features to their programs.

It will be increasingly up to the software developers to provide the speed, not Intel or AMD, and those who don’t heed the new regime will likely become victims of it.

Ed


Be the first to comment

Leave a Reply