Intel's new roadmaps have been the talk of the town for the last few days due to the negative press swirling around cancelation of Tejas and Jayhawk (Pentium 4 successors). Fortunately, with the Intel Channel Conferences occurring right now, we were able to sneak a peak at some additional technology Intel has planned, as well as update our roadmaps for the new processors.

Intel's largest addition on the horizon? As the topic suggests, 1066FSB has been added to the Q3 roadmap, but for only two processors; the 3.46GHz P4EE and the 3.73GHz Prescott Pentium 4. The expected update on FSB has been long overdue; many 865PE and 875P solutions already claim support for 1GHz+ front side bus. Overclockers should be fairly familiar with this.

However, despite the suggestion that new 1066FSB processors are on the way from Intel, the next generation chipsets do not immediately support the new bus speed. ATI's roadmap for Intel already hints at 1GHz FSB, but SiS and others have been fairly quiet. Since it will be at least Q2'05 before the next Intel chipset debuts (Lakeport), Intel has added another chipset to its 2004 desktop lineup. This new chipset, dubbed 925XE, is exactly the same as the 925X chipset with the addition of 1066FSB support. It may be quite probable that 925X and 915P are overclockable to 1066FSB speeds, but as of today the 925XE is the only officially announced core logic.

Not only does the Prescott receive an extra boost in FSB, but Intel has slated a new increase in cache size for the Pentium 4 (not P4EE) family. The first new processor (Intel 720) will carry a 2MB L2 cache! Recall that Dothan (Pentium M) also incorporate 2MB L2 cache. With the longer pipelines on the Prescott and Dothan processors, the additional L2 cache has proven invaluable to keeping the processors competitve. The P4EE with 1066FSB will only have 2MB L3 cache, but the L2 cache has not been disclosed. Whether or not the newest P4EE is Prescott or Northwood based remains a mystery, but since most P4EEs have been based off the Gallitin (Xeon) cores in the past, it is probable that the processor is again Gallitin or Nocona based.

Pentium Roadmap (5xx)
CPU Manufacturing Process Front Side Bus L2 Cache Size Product Name
Pentium 4 EE 3.46GHz ??? 1066MHz ??? ???
Pentium 4 EE 3.4GHz 130nm 800MHz 512KB ???
Pentium 4 EE 3.2GHz 130nm 800MHz 512KB ???
Pentium 4 4.0GHz 90nm 800MHz 1MB 580
Pentium 4 3.8GHz 90nm 800MHz 1MB 570
Pentium 4 3.6GHz 90nm 800MHz 1MB 560
Pentium 4 M3.6GHz 90nm 533MHz 1MB 558
Pentium 4 3.46GHz 90nm 800MHz 1MB 552
Pentium 4 3.4GHz 90nm 800MHz 1MB 550
Pentium 4 3.2GHz 90nm 800MHz 1MB 540
Pentium 4 M 3.2GHz 90nm 533MHz 1MB 538
Pentium 4 M 3.06GHz 90nm 533MHz 1MB 532
Pentium 4 3.0GHz 90nm 800MHz 1MB 530
Pentium 4 2.8GHz 90nm 800MHz 1MB 520
Pentium 4 M 2.8GHz 90nm 533MHz 1MB 518

The only real addition to the mirange desktop roadmap is the P4EE.

Pentium Roadmap (7xx)
CPU Manufacturing Process Front Side Bus L2 Cache Size Product Name
Pentium M 2.13GHz 90nm 533MHz 2MB 770
Pentium M 2.0GHz 130nm 533MHz 2MB 760
Pentium M LV 1.50GHz 130nm 400MHz 2MB 758
Pentium M 2.0GHz 90nm 400MHz 2MB 755
Pentium M ULV 1.2GHz 90nm 400MHz 2MB 753
Pentium M 1.86GHz 90nm 533MHz 2MB 750
Pentium M 1.8GHz 90nm 400MHz 2MB 745
Pentium M 1.73GHz 90nm 533MHz 2MB 740
Pentium M LV 1.40GHz 90nm 400MHz 2MB 738
Pentium M 1.70GHz 90nm 400MHz 2MB 735
Pentium M ULV 1.10GHz 90nm 400MHz 2MB 733
Pentium M 1.60GHz 90nm 533MHz 2MB 730
Pentium M 1.60GHz 90nm 400MHz 2MB 725
Pentium 4 3.73GHz 90nm 1066MHz 2MB 720
Pentium M LV 1.30GHz 90nm 400MHz 1MB 718
Pentium M 1.50GHz 90nm 400MHz 2MB 715
Pentium M ULV 1.10GHz 90nm 400MHz 1MB 713

Notice there have been some significant naming changes since the last roadmap we published. It seems the Ultra Low Voltage (ULV) and Low Voltage (LV) naming convensions have flip flopped. The LV processors are now regarded as the "higher" numbered chips. Even though the product name is not supposed to denote performance, it seems the move to rate the LV processors higher than the ULV processors (a reversal on the original naming) seems to suggest otherwise.

You may notice the Intel 720 is listed in the 7xx family of processors and that these processors all share something in common; Banias or Dothan core.

Below are you can see our updated Celeron roadmap as well. Celeron D technically launches in June, but we have already seen multiple samples from retail merchants.

Celeron Roadmap (3xx)
CPU Manufacturing Process Front Side Bus L2 Cache Size Product Name
Celeron M ULV 1GHz 90nm 400MHz 512KB 373
Celeron M 1.5GHz 90nm 400MHz 1MB 370
Celeron M 1.4GHz 90nm 400MHz 1MB 360
Celeron M ULV 900MHz 130nm 400MHz 512KB 353
Celeron M 1.3GHz 90nm 400MHz 1MB 350
Celeron 3.2GHz 90nm 533MHz 256KB 350
Celeron 3.06GHz 90nm 533MHz 256KB 345
Celeron M 1.5GHz 130nm 400MHz 512KB 340
Celeron 2.93GHz 90nm 533MHz 256KB 340
Celeron 2.8GHz 90nm 533MHz 256KB 335
Celeron M 1.4GHz 130nm 400MHz 512KB 330
Celeron 2.66GHz 90nm 533MHz 256KB 330
Celeron 2.53GHz 90nm 533MHz 256KB 325
Celeron M 1.3GHz 130nm 400MHz 512KB 320
Comments Locked


View All Comments

  • OCedHrt - Sunday, May 30, 2004 - link

    Nice to see that the Pentium 735 seems to be now available.
  • KristopherKubicki - Tuesday, May 18, 2004 - link

    New AMD and Intel info this week....

  • stephenbrooks - Sunday, May 16, 2004 - link

    Good to see another Intel CPU roadmap from the Twilight Zone. Personally, I'd really like to get my hands on that 3.73GHz/2MB/FSB266x4 part. Who knows what they're up to with that? I'd expect the reason it's in the 7xx series is that with the doubled cache, its die area and price would double and be comparable with others in the 7xx series.
  • bhtooefr - Thursday, May 13, 2004 - link

    What I've heard is this:

    7xx - PM, P4EE
    5xx - P4, P4M
    3xx - Celery, CM
  • KristopherKubicki - Tuesday, May 11, 2004 - link

    Third week of june.

  • JackHawksmoor - Tuesday, May 11, 2004 - link

    Yeah, this is getting really weird...

    Do we have a time table for the release of the new P4 chips? We've got their names now, but no timetable...
  • mlittl3 - Tuesday, May 11, 2004 - link

    I think the numbering scheme is equivalent to how badly Intel wants to sell the chip.

    7xx - we really want people to buy these chips
    5xx - we sort of want people to buy these chips
    3xx - we really don't want people to buy these chips

    I hardly believe that the Pentium M @ 1.6 GHz beats a P4EE at 3.4 GHz. I could be wrong. If so, someone send me a link showing the banias or dothan core beating an extreme edition CPU in a decent number of benchmarks.
  • bobsmith1492 - Tuesday, May 11, 2004 - link

    So, for a '1066' bus, the actual bus clock would be 266.5 (asuming a quad-pumped thingamajigger), and synchronized DDR would run at 266.5*2, or 533. Isn't this one of the first official speeds for DDR2? So, it seems like this bus speed would require DDR2.
  • Runamile - Tuesday, May 11, 2004 - link

    Is it just me, or is this a whole lot more choices then there should be? I counted 46 processors. 46? whats wrong with about 6 or so for each class? This is going to cause way too much confusion for the common consumer. Basic marketing does state having 3 distict grades/classes of quality, but having 15 subclass inside each?

    Also, the fact does remain that the fastest of the 3xx can beat the slowest 5xx, and the fastest 5xx can beat the slowest 7xx (I am speculating here). Isn't that also going to cause confusion? It appears that when Intel announced their new naming scheme months ago they were counting on cutting out their P4 lines by looking at how almost all the 5xx series was filled up at the time. I hope that line is cut out pretty soon, to aleviate the amount of choices. Too many choices is frusterating to consumers. The mainstream consumer just want a computer, they don't care what the exact specs or model number of their CPU is. Telling a person they should get a 735 over a 733 isn't going to mean crap to them. They just want to know what is fastest or cheapest. Period.
  • Pumpkinierre - Tuesday, May 11, 2004 - link

    A suggestion the new Prescott 720 is a dual core Pentium M 750. 2x1.86MHz is 3.73 (allowing for the 3rd decimal) and 2x533MHz FSB is 1066. Its been strangely lumped in with the P-M mobile CPUs and its got a weird speed. Also, Its got 2Mb of cache same as the P-Ms and even for a dual core that's good enough. One of the other posts (Sudhan?) said there could be two types of Prescott and Intel said they were moving to dual core and cooler chips (most probably Pentium-Ms). Maybe they are really panicking at Intel!

Log in

Don't have an account? Sign up now