• Tosan25@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    10 months ago

    Celerons went socket before the P3 did, and the and were quite competitive with the on-die cache so they had to be crippled with less cache and a lower FSB so as not to cannibalize the much higher prices P3s. Katmai P3s also had the same back side L2 cache that ran at half speed into Coppermine.

    Intel sold Coppermine P3s at both 100 and 133 FSB, so putting a Celery at 100 FSB would have made it too performance competitive with the 100 FSB P3s, making people but the cheaper chip.

    The 100 FSB Coppermine P3s overclocked very well, with the 700 MHz being the sweet spot. Most people could bump the FSB to 133 to get a 933 MHz chip with little effort right out of the box. I did.

    Tualatin Celeries and P3s did even better, thanks to a process shrink to 130nm. The only real difference was the FSB at that point (100 vs 133) as they both had the same L2 cache, except the P3-S which has 512K.

    Tualatin Celerons could overclock like crazy. It was my best overclock until Core 2 Duo. Mine got to 150 FSB. We had a competition at work to see who’s system could perform the best, with previously agreed upon benchmarks. My Celery destroyed several Willy P4s, much to the shock and dismay of my co-workers. 😁 The only thing one of them could say was “yeah, well yours is a Celeron. Mine’s a P4 so it’s better.” Of course, if it was so great, I shouldn’t have owned it with a lowly Celery. 😁

    The P3 was a damn good design. It’s a shame Intel favored NetBurst for so long instead of developing the P4. I think they would have been more competitive against AMD despite the clock speed wars of the time.