RealTechNews.com - Independent Tech

Monday, November 01, 2004

It's the End of the (Pentium) 4 As We Know It
Two weeks from now, Intel will crank up the P4 one last time to 3.8GHz with 1MB of L2 cache. This marks the end of Intel's single core CPU line and a move now towards dual core processors that will rely on tweaks beyond mere horsepower.

Bill of course, has called the move nothing but horse-pucky (is that a word?) and I'll let him weigh in here later on the issue. But for now, it looks like 3.8 is as high the 4 is going to go and that just doesn't add up to me one bit.
# Permalink Posted at 9:46 PM 4 comments Email this Link

Comments on this Item:

Correct me if I'm wrong, but the real world difference between a 3.8 GHz and a 4.0 GHz CPU after you factor in bus bottlnecks etc. is going to be in the magnitude of splitting hairs. Sure, a benchmark might see it, but in every day life you're not likely to.

And if you DO do something computationally intensive enough that the difference actually matters, wouldn't you be better off with a multi-CPU system (server or workstation) anyway? A 2x3.x GHz Xeon box sounds a lot faster than a single 4 GHz to me....

# posted by Anonymous 8:36 AM

But that's the same old story and to me it's not an either/or issue. I want faster horsepower and faster components and secondary systems. All I'm hearing is that they went as fast as they can.

# posted by Alice 11:50 AM

I think it's more a case of "as fast as they can reliably by ratcheting up the clock speed."

Case in point, take a look at the P-M line. It doesn't have anything approaching the clock speed of the P4 line, but in real world computing it's not far off in terms of "I have to do X, when can I have it done." What it really comes down to is, it looks like making a CPU more powerful by increasing the clock speed has hit the diminishing returns wall, and there are other more productive means to get there from here.

# posted by Anonymous 1:11 PM

Alice like many computer people you lack experience with electronic design. The Prescot core HEATS UP!!! Couple that with narrow a thermo-dynamic range relative to output and you have a package that could be unstable at higher speeds.

# posted by Anonymous 12:04 PM

Post a Comment