Successor to Pentium 4?
 

[url]http://news.yahoo.com/s/ap/20050812/ap_on_hi_te/intel_chip_changes;_ylt=AgXBFypsN8STZmUcyztWEu2s0NUE;_ylu=X3oDMTA3cjE0b2MwBHNlYwM3Mzg-[/url]

"The world's largest chip maker will announce the architecture this month at a conference in San Francisco, spokesman Bill Calder said Thursday."

Edit: More specific information here:

[url]http://news.zdnet.com/2100-9584_22-5829313.html[/url]

Unless Intel makes major changes to the Pentium M FPU, this will be a giant step BACKWARDS for prime hunting. AMD64 chips will be the clear CPU of choice.

[QUOTE=Prime95]Unless Intel makes major changes to the Pentium M FPU, this will be a giant step BACKWARDS for prime hunting. AMD64 chips will be the clear CPU of choice.[/QUOTE]

It will?! :w00t: :sad: :no: :down:

Why is that?

(Also, what does FPU stand for? Please excuse my ignorance).

FPU = Floating-point Unit (the guts of the processor that perform floating point [as opposed to integer] math operations)

Lower power consumption and mobile use are driving many of major CPU manufcaturer decisions these days - why do you think Apple decided to ditch the G5? (One irony there being, much of the G5's tremendous FPU capability was never accessible to most of Apple's customers due to Apple's mediocre compiler technology - they basically never moved significantly beyond plain-ol' gcc.)

[QUOTE=ewmayer]Lower power consumption and mobile use are driving many of major CPU manufcaturer decisions these days - why do you think Apple decided to ditch the G5? (One irony there being, much of the G5's tremendous FPU capability was never accessible to most of Apple's customers due to Apple's mediocre compiler technology - they basically never moved significantly beyond plain-ol' gcc.)[/QUOTE]

I have to disagree with you here. Apple went to great lengths to improve gcc. gcc 4 is much better at optimizing than gcc 3. In comparison to IBM's xlc, gcc 4 actually optimizes some FFT implementations better than xlc (gcc 3 didn't optimize anywhere near as well). There are still some FFT implementations, such as your with Mlucas, that still optimize better with xlc than gcc 4, but it has come a long ways.

So that answers my second question.

But what about my first? Why would this new processor be slower than P4?

if they dont edit the pentium m unit yes it will be slower.

[QUOTE=jinydu]Why would this new processor be slower than P4?[/QUOTE]Pentium M uses a slower memory access architecture than the Netburst architecture used in the fastest P4s now.

From the second article you linked in your initial posting:
[quote]Current Pentium 4 desktop chips and Xeon server chips come from the NetBurst architecture.

The power consumption of the NetBurst chips, however, has created problems for PC makers and limited performance gains.[/quote]

That memory architecture difference is in addition to what George referred to about FPUs (about which I know no details).

lets just hope that Intel has a few cards up their sleeves so they can finally kick some AMD ass!!! I cant wait until the announcement at the forum.