Anand has a prescott?

[rgreen83]

Originally posted by: MadRat
90nm, really? The internal bandwidth of the Prescott should set it way apart from the older Pentium-4. They probably got the thing crippled, Celeron-style, as far as connections and associativity to the internal caches.

....AND there it goes... way off to left FIELD!

Electrons will only move through the cpu as fast as they are clocked, regardless of manufacturing size (withholding stalls). The 90nm wont by itself make the prescott any faster, but the 90nm and strained silicon enable intel to crank up the clock speed faster. Thats is the advantage of 90nm. And as for crippling it like the celeron... Why would they do that to their next flagship cpu?

There is though, as I understand it, a bottleneck that can occur when the transistor gates get too narrow, but I dont think this becomes a problem until around 45nm process, and that will be solved by the time they get there with things like triple gates and such.

 

[rgreen83]

I wonder if we tipped someone off? look here

 

[MadRat]

Who is to say they haven't saved some of the starch for Prescott until the initial launch? Did they not disable functions in the original P4 because the technology was deemed too immature? Did the P4-based Celeron not start out as a full-blown Pentium 4 with functionality and internal pathways to the internal caches not crippled? If they could get engineering samples out to reviewers and leave an impression of blah with the early reviewers then whats the harm? If the product at launch suddenly becomes a "WOW!" by enabling the hidden features then Intel only enhances its product launch.

 

[rgreen83]

I understand what you are saying but remember, this is intel, they wouldnt share the limelight with anyone if the fate of the universe swung in the balance.
That was the reason for the p4 emergency edition, and if they can help it. Hyperthreading was in all northwod cores, but wasnt enabled until the 3.06B came out because of problems, but would they have enabled it sooner if possible? you bet your pants they would have. That may be some of the situation here, problems are keeping the prescott from taking off. And I thought a celeron was a northwood core with 3/4 of the cache simply disabled, as with the duron? I know wingznut knows, but i wont ask him.

 

[DivideBYZero]

Originally posted by: rgreen83
I wonder if we tipped someone off? look here

That Charlie Demerjian is a fool. I sincerly hope he's signs up an account here, rather than just lurking. The flaming he would get would be so worth it.

 

[OddTSi]

Originally posted by: rgreen83
That may be some of the situation here, problems are keeping the prescott from taking off.

Maybe I missed something, but isn't Prescott still on schedule for its release? Why is it that everyone is saying that it's having problems and that it's being delayed? IIRC it was always scheduled as a Q4 2003 release, and it still on that schedule apparently.

 

[redpriest_]

I wasn't comparing it to a A64, FYI. I was comparing it to a p4c's -- it gets owned by the P4EE, perhaps lending credibility to the claim that in games, the P4 EE is faster than a prescott.

 

[glugglug]

Looking at those Unreal benches, The 2.8GHz Prescott is slower than a P4 3.0C. Maybe an even match for a P4 2.8C, perhaps even a little slower than that.

What are the improvements in Prescott vs. P4 besides 90nm process and PNI? Do applications that are not PNI aware see no benefit at all?

 

[rgreen83]

Originally posted by: OddTSi

Originally posted by: rgreen83
That may be some of the situation here, problems are keeping the prescott from taking off.

Maybe I missed something, but isn't Prescott still on schedule for its release? Why is it that everyone is saying that it's having problems and that it's being delayed? IIRC it was always scheduled as a Q4 2003 release, and it still on that schedule apparently.

Actually the prescott was originally scheduled as q2 2003. But this still isnt as bad as AMD being late by two years with the K8.

 

[Wingznut]

Originally posted by: rgreen83
Actually the prescott was originally scheduled as q2 2003. But this still isnt as bad as AMD being late by two years with the K8.

Are you sure about that? For as long as I can remember, Intel has always been shooting for a "H2 '03" release.

I did a little searching, and found this AnandTech article, dated 03/12/02"

Just two weeks ago Intel officially introduced the successor to the Pentium 4 under the codename Prescott. To be released in the second half of 2003, Prescott will be Intel's first 0.09-micron (90nm) CPU that will feature the following...

And then here is the official press release announcing Prescott, dated 02/27/02:

This platform is based on a processor with the code name Prescott. Prescott will be shipping in the second half of next year...

 

[jbond04]

Originally posted by: rgreen83

Actually the prescott was originally scheduled as q2 2003. But this still isnt as bad as AMD being late by two years with the K8.

Like Wingznut said, I'm pretty sure that Prescott has always been slated for a H2 2003 launch...although I do wish they would hurry up with it. 😉

 

[Duvie]

from what I heard from a reliable source that the prescott has 4 steppings already....Also that it is rumored around Intel that the prescott does have some magic pill or some sort of feature that only the prescott team knows about. What could this be and maybe Anand has to disbale it in this early testing not to let cat out of the bag or thatmaybe only the newer boards will activate it. I get a feeling it could be big.

On the other hand I think it is clear as an ocer the prescott may not be for me since the heat issues are real and will really limit the gains many of us p4 users are use to.

I will likely just pick up a p4c here in a short while and see if the steppings get a bit better and can insure me 3.6ghz on an oc.

Maybe even a A64 after the socket change and ther multimedia numbers get better...otherwise it is almost a step down or not enough a an increase from my 2.4b@3.24ghz now.

 

[rgreen83]

Originally posted by: Wingznut

Originally posted by: rgreen83
Actually the prescott was originally scheduled as q2 2003. But this still isnt as bad as AMD being late by two years with the K8.

Are you sure about that? For as long as I can remember, Intel has always been shooting for a "H2 '03" release.

I did a little searching, and found this AnandTech article, dated 03/12/02"

Just two weeks ago Intel officially introduced the successor to the Pentium 4 under the codename Prescott. To be released in the second half of 2003, Prescott will be Intel's first 0.09-micron (90nm) CPU that will feature the following...

And then here is the official press release announcing Prescott, dated 02/27/02:

This platform is based on a processor with the code name Prescott. Prescott will be shipping in the second half of next year...

I stand corrected. Thanks wingznut, I think I mistook "H2" 2003 for "Q3" 2003.

 

[Wingznut]

Originally posted by: rgreen83
I stand corrected. Thanks wingznut, I think I mistook "H2" 2003 for "Q3" 2003.

Well, in all fairness... You've probably read "Prescott delayed" a lot at sites such as theinquirer or theregister, or even msg boards. I'll bet that's where the impression came from.

 

[rgreen83]

I think I just spend too much time "anticipating". For that reason everything seems a delayed these days. Maybe someday we will have the large jumps like we used to see. I am constantly disappointed when "Hyper this" and "Extreme that" fail to improve more than 5-10% over yesterdays technology. The golden age seems to be over, every upgrade seems to be "more of the same" instead of true innovation. Hell AMD might fail with their integrated mem controller and 64-bit, but thank god someone at least tried something new.