-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
why did intel make release the prescott core?
- Thread starter l3ored
- Start date
- Joined
- Apr 26, 2004
- Location
- The Netherlands
As a company as Intel, you never know exactly what gains and what problems you are going to get from a new CPU, and probably even less from a new process shrink like prescott. (130 --> 90nm). This time it turned out very bad, the 90nm has some very weird power issues. If they wouldnt have been there from the beginning, we would have been with the 4Ghz prescott now (keyword: *would*).
The prescott scales better then northwood, on 3 Ghz the NW is better, on 4 GHz the prescott. Go beyond and the gap increases. Now even AMD is said to face 90nm heat issues (tho less then Intel), they both 'agreed' that dual core is the way to go.
The 1 MB cache of prescott gives it a nice bump in some apps like SETI and CAD, and when the PNI (prescott new instructions) get implemented and used by software makers, it will perform even better.
The prescott scales better then northwood, on 3 Ghz the NW is better, on 4 GHz the prescott. Go beyond and the gap increases. Now even AMD is said to face 90nm heat issues (tho less then Intel), they both 'agreed' that dual core is the way to go.
The 1 MB cache of prescott gives it a nice bump in some apps like SETI and CAD, and when the PNI (prescott new instructions) get implemented and used by software makers, it will perform even better.
- Joined
- Jan 23, 2003
- Location
- England
The reason Intel are selling lower clocked prescotts is because although they might be slower, and they might run hotter, they are interested in production costs, for them, 90nm is cheaper to create than 120nm, simply because of die space, the prescott is a smaller core than the northwood, so they can produce them for less, and make more money.
- Joined
- Jun 19, 2002
The theory is that by shrinking to 90 nm they will be able to clock them higher in the future than the current 130 nm Northwoods, although unfortunately, there's plenty of heat issues.
- Joined
- Jul 23, 2003
- Location
- Berkeley, CA
Don't forget, when the P4 was first released it was spanked by P3's and Athlons. Hell, even higher Durons were beating the early P4s, and those ran pretty hotly if my memory serves. The 90nm process needs refinement. Since they're going with the P-M architecture, i'd like to see the experience gained from grappling with the issues of the 90nm process put to good use in the future.
- Joined
- Jun 16, 2003
- Location
- Cranberry Twp. PA
forgive me if this sounds stupid, but wouldn't it make sense to space out the transisitors? Even a nanometer or two wouldn't make Intel (or amd) lose that much money in wafers. With the transistors spaced out, there would be more area to spread the heat, and that might net maybe a 1-2C loss? =D
I don't think heat is especially the problem here. The heat comes from leakage in the transistos, and spacing would'nt make a differance for that. Currently it seems like just about everyone is having problems with 90nm, but they will be sorted out sooner or later... (maby) 
- Joined
- Jan 2, 2004
Yeah, but Intel doubled the amount of transitors in P4 Prescott, when compared to Northwood.
The doubling of both the L1 D-cache and L2 cache cannot account for all ~130 million transitors.
The doubling of both the L1 D-cache and L2 cache cannot account for all ~130 million transitors.
- Joined
- May 1, 2002
- Location
- Tampa, Florida. USA
the p3 and athlon "spanked the p4"??
I dont think so..
while the p3 and athlons were around 1 ghz, and the p4 was introduced at 1.5 ghz with dual channel rambus
there was no spanking...
if the p4 was introduced at 1.0 ghz, it would have lost
but it wasnt...
I dont think so..
while the p3 and athlons were around 1 ghz, and the p4 was introduced at 1.5 ghz with dual channel rambus
there was no spanking...
if the p4 was introduced at 1.0 ghz, it would have lost
but it wasnt...
k_r_i_p_l_e_r
Disabled
- Joined
- May 16, 2004
- Location
- Jamaica, NY
the p4 was introduced a 1.0 ghz..... williamette
- Joined
- Apr 26, 2004
- Location
- The Netherlands
AFAIK the first P4s were 1.2Ghz, and were *seriously* kicked by P3 1133Mhz, and the athlon 1 Ghz. Those P4's werent too popular
Only when they reached 1.8Ghz, they got a little advantage over the older CPU's, and they got really serious with Northwood.The P3 seems to be one of intels best chips until now. A long time ago, there was a benchmark somewhere that put a 1.7Ghz clocked P4 vs. a P3 1133 @ 1.7Ghz, and the P4 was beaten by like 20%, except in memory bandwidth. If intel wouldve shrunken the die size of the P3, and added some extra memory bandwidth, they could have saved themselves alot of R&D on prescott
Last edited:
- Joined
- Jan 2, 2004
nipster said:
Bzzzzt, wrong.
My Tbird 1000 spanked the Willamette P4 1.4 that we had in the house for a while.
Higher clock speeds are not necessarily indicative of higher performance and can only be used as a comparative factor between processors of the same core design.
- Joined
- Jul 23, 2003
- Location
- Berkeley, CA
Captain Newbie said:
Yep. Don't you all remember the angry articles? People even sued / are suing because they Intel's marketing campaign for the P4 was misleading due to the fact that it was *slower* than the Pentium 3. (rumour?)
Fast420A said:
Need it for what? Probably it helps a bit to counter the lost performance from the longer pipeline, but if it does'nt help performance (over the northwood) then it's not really that needed.
Ofcourse once ddr II comes out with higher latencys it can help.
k_r_i_p_l_e_r
Disabled
- Joined
- May 16, 2004
- Location
- Jamaica, NY
so some p3s have higher clock speed than some p4s?
- Joined
- Jul 17, 2002
No. You totally missed the point. The clockspeed doesn't mean anything when you're comparing actual performance. The older and slower clocked pentium 3's in the 1.1ghz range matched or beat the early pentium 4's(socket 423, not the current socket 478) in almost every benchmark except memory bandwidth. That was the case for months.
What made it even more ridiculous was that intel wound up releasing pentium 3's(tualatin iirc) clocked up to 1.4ghz, that probably slaughtered the early p4's.(they also currently cost like twice as much too)
What made it even more ridiculous was that intel wound up releasing pentium 3's(tualatin iirc) clocked up to 1.4ghz, that probably slaughtered the early p4's.(they also currently cost like twice as much too)
Similar threads
