old school overclocking

ScuffScuff Southwestern, Pennsylvania
edited February 2005 in Hardware
Iv' e been messing around with a p2 350 on a few motherboards and a celeron 766 on a different mobo. It seems the 350 actually outperforms the 733 when i raise the fsb to like 112 on the 350. Does this sound correct?
And how important is the 512 cache over the 128 of the celeron? what difference does this create?

Comments

  • mmonninmmonnin Centreville, VA
    edited February 2005
    Cache makes a huge difference in performance. Celerons suck because they dont have as much cache as their couterpart. Celerons are the same as P4s or P3s with less cache.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited February 2005
    mmonnin wrote:
    Celerons suck because they dont have as much cache as their couterpart.

    Correction: MOST celerons suck. The tualatin core Celerons were not bad CPUs at all. They had the same amount of cache as the regular tualatin core PIIIs (the PIII-S had 512k but the regular Tualatin PIII had 256k)... the only difference was the 100MHz FSB vs. the 133 of the PIII. Despite its 100MHz SDR FSB, the Tualatin Celeron was about as fast as an Athlon/Athlon XP clocked something like 100MHz lower than the Celeron. The PIII-S was functionally equivalent to the Athlon/Athlon XP, i.e. a 1.4GHz PIII-S performs effectively the same as a 1.4GHz Thunderbird or Athlon XP 1600+. And it did it with <1/2 the heat output of the Athlon. When I had my MSI Pro266TD-LR running, it had a 1.3GHz Celeron in it... it got used more often than any of the other desktops (which at the time would've been the dual athlon, a single athlon of some sort, and I think maybe the P4 :scratch: ) because it was absolutely silent (the filter on my aquarium was louder) and more than fast enough for most of the work I do, including some gaming.

    Also, the first Celerons were based on the PII, not the PIII, and I believe the very first ones had no cache at all. Or no L2 cache. Or something like that. :shakehead
  • profdlpprofdlp The Holy City Of Westlake, Ohio
    edited February 2005
    Intel has a long history of releasing chips of all flavors with different sized caches, usually as a way to better fit different niches in the business market. An example would be a lower-cache model offered as a workstation PC in an environment where CPU performance is not a big deal.

    If you've got an office full of employees who are spending their entire day responding to email and writing out reports in Word format you don't need a big hairy processor. I used to have a law firm as a client back in Virginia who had about forty lawyers and twenty secretaries. (You can probably guess which of the two groups had to work twice as hard as the other.) ;)

    Sixty computers in all, (not counting a couple servers), none of which needed a vast amount of processing power - they spent 95% of their time doing the two tasks I listed above.

    One of the most interesting cache-related events was Intel's Celeron 300A. About the time that AMD was starting to look like a viable alternative, Intel released the Celeron line, hoping to hang on to the customers who were shopping based on price alone - who were buying a lot of K6-2's. The early Celeron's came out with NO cache at all and had their butts whipped pretty bad. Intel then came out with the 300A, with a 128kb cache (compared to the PII 512kb) but with the huge advantage of having the cache on-die, meaning it ran at the full speed of the CPU. PII processors ran at the speed of the FSB.
    The Intel Celeron processor in the S.E.P. package at 400 MHz, 366 MHz, 333 MHz and 300A MHz have an on-die 128-KB L2 cache and the Intel Celeron processor at 300 MHz and 266 MHz have no L2 cache. The "A" is added to the "300A MHz" of the Intel Celeron processor at 300A MHz to distinguish it as having an L2 cache while the Intel Celeron processor at 300 MHz does not have an L2 cache.
    Source - Intel

    Enthusiasts flocked to the 300A in droves for two basic reasons. 128kb worth of cache running at 300MHz did about as well as 512kb of cache running at 66MHz or 100MHz. The second reason was that by moving one jumper on your MB to enable a 100MHz FSB (the celery 300A was designed for a 66MHz board) you suddenly had a CPU running at 450MHz which would perform as well as the top PII's of the day. I remember that most of the guys in the shop where I worked at the time (1999) put together just such a system. You don't see a whole lot CPU's running at 150% of their rated speed anymore. :cool:
Sign In or Register to comment.