Dual core PresHot, what's the point?

MJOMJO Denmark New
edited June 2004 in Hardware
I just saw this at another site.
Intel Corporation may release dual-core Pentium 4 “Prescott” processors in late 2005, a report over Geek.com web-site claims. If the information is correct, the roadmap of the world’s largest manufacturer of central processing units gets completely reshuffled once again.

Dual-Core Prescott Spotted

At the Intel/PC Magazine-sponsored Technology for Business Today seminar in Washington, D.C., Intel representatives discussed the present and future of computers touching upon the modern and next-generation microprocessors. The firm’s officials reiterated the company’s plans to issue dual-core processors in 2005 and even shed some light on the technical information about the chips. Apparently, the desktop processors will continue to utilize the NetBurst architecture and at this point such central processing units are referred as dual-core Prescott microprocessors. Mobile dual-core chips will have architecture similar to that of the Pentium M products available today. Such products may also find themselves in desktops, though, the premier performance will be offered only by dual-core NetBurst products.

http://www.xbitlabs.com/news/cpu/display/20040610151158.html

Why on earth is intel working on a dual core PresHot?
I thought they more or less gave up on PresHot architecture when they scrapped the Tejas.
They are supposed to be working on a desktop Pentium-M and a dual core Pentium-M, not a dual core PresHot?
According to the news they are working on a dual core Pentium-M, but it is only for the notebooks.

It seems that this dual core PresHot will use the same socket as the recently scrapped Tejas.

But I wonder, wouldn't two cores use double wattage(in principle)?
2x103W=206W, guess all PresHot fanboys needs to move north of the arctic circle and/or use phase changing cooling. ;)

Note: I dunno whether the news is correct I'm just relaying information.

Comments

  • mmonninmmonnin Centreville, VA
    edited June 2004
    No it wouldnt use double wattage. More like 25-40% more I would think. Depends on the number of transistors. They pprolly wont double everything on the core so it wont double.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited June 2004
    Point is, Prescott is likely to become a mid-range CPU position in 2005. With two cores, why does the base GHz rate have to be so hugely high (and power and heat dissipation hungry) per core??? Answer, it does NOT have to be so high with a true dual-core to get performance good enough for a midrange. For high-end, try Dual-Core Xeons.
  • muddocktormuddocktor
    edited June 2004
    How about for high end having dual core Dothans. :D
  • qparadoxqparadox Vancouver, BC
    edited June 2004
    On the Presshot cache takes up~60% of the die IIRC. If they take out any of that their performance is screwed. Most of the other circuitry is not redudant, maybe they could have some of the instruction decoders and such combined, and the external interfaces. But likely 90%+ of the chip would be replicated. The space (and heat) savings comes from the switch to a 65nm process (theoretically) which means the double presshots will *only* draw maybe 30% more combined than the current presshot. Its still a lot but for an essentially dual processor machine its well within reason.
  • muddocktormuddocktor
    edited June 2004
    But qp, Intel (and it also looks like IBM too, with watercooling needed for the new 2.5GHz Mac G5) is already having current leakage problems with 90 nanometer so why should 65 nanometer be even as efficient as the current presshot? Unless someone finds some kind of breakthrough in technology, I think that all manufacturers will find 65 nanometer fabrication even harder than they are finding 90 nanometer fabrication. Just remember, presshot's power consumption went up compared to Northwood even though it's based on a .09 process compared to Northwood's .13 process.

    Unless there is some kind of breakthrough in technology, I can't see dual core P4 as any kind of choice at present clock speeds.
  • mmonninmmonnin Centreville, VA
    edited June 2004
    Its called SOI. There are also other techniques that will hopefully keep the leakage to a minimum.
  • muddocktormuddocktor
    edited June 2004
    Marc, the G5's are using SOI and they are having problems too.

    Anyways, I read over at ARS Technica that the Inel fool talking this crap is a marketing manager, who is probably talking out of his ass for what he really knows.

    Here's a quote from the last line in that story:
    Update: Intel has said that the company has provided no such specifics publicly. 60% of the time, that means there's a little truth to the story. Of course, which part is true we're not sure.
  • mmonninmmonnin Centreville, VA
    edited June 2004
    I thought the Opterons and such were using it as well. They run pretty cool.

    Ive also head that the probe is wrong on for the Prescotts. Running it like 25C hotter than it actually is.
  • muddocktormuddocktor
    edited June 2004
    SOI is working well on present Hammer chips, but they are still all .13 micron. No one actually knows how they are running on .09 besides AMD right now. The one thing I read about .09 Hammers was at overclockers.com, which said that AMD might be having some heat/power problems too. Here is a link to the article where they are talking about this.
Sign In or Register to comment.