Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

AMD Model Numbering System...Baffeling?

Options
  • 30-05-2005 6:06pm
    #1
    Moderators Posts: 5,555 ✭✭✭


    Right on Socket 754 AMD had 2 versions of the 3200 out.

    A 2.0Ghz 1MB Cache and a 2.2Ghz with 512kb cache.
    I have a 3400 socket 754 at 2.2Ghz with 1MB cache.

    The 3400 is obviously superior to either of those chips thanks to the extra cache. Its priced dearer.

    The final 64bit chip from AMD is the 3700 at 2.4Ghz with 1MB cache.
    Now consider there nearest (up till recently) counter parts on socket 939 the 3500 and 3800. Now including the latest cores and core revisions to these chips (like venice) and weighing in the dual channel support via single channel and the fact that the 939 chips have only 512kb cache. How do these chips actually compare.

    Next we the new San Degio 3700 which has 2.2Ghz and 1MB cache and is cheaper than the Venice 3800 with 2.4Ghz 512 kb cahce. I always belived the extra cache was more benefical than a mere 200mhz clock speed increase and that the chip with more cache was marginally faster.

    I'm wondering what the actually performance difference between these chips is?


Comments

  • Registered Users Posts: 16,608 ✭✭✭✭astrofool


    they're all relative really, and it depends on the benchmark, sometimes extra cache makes a big diff, sometimes dual channels, sometimes neither.

    The early 3000+'s were almost the same as the 3200+, even though they were half the cache.


  • Registered Users Posts: 4,003 ✭✭✭rsynnott


    It's worth noting that at least the San Diego and possibly also the Venice support SSE3, and have pipeline changes.


  • Registered Users Posts: 15,815 ✭✭✭✭po0k


    It's just a modelling scheme to differentiate chips in each socket family, Skt 754 and skt 939 shouldn't be compared (strictly). Course, it won't stop people doing it smile.gif


  • Registered Users Posts: 13,016 ✭✭✭✭vibe666


    people still need to know what they're buying though without having to get out a calculator and slide rule!

    i think most (non-uber-geeky) people would just go on the speed numerical (3000+, 3400+ etc) and the price and work out the rest from there depending on their budgets.


  • Registered Users Posts: 144 ✭✭r00t69


    I think extra Cache is usefull in certain cicrumstances but extra clock speed is always useful, thats why I popped for a 3800+ V over the new 3700+ SD.

    This might be slighly over simplified :rolleyes:


  • Advertisement
  • Registered Users Posts: 16,608 ✭✭✭✭astrofool


    vibe666 wrote:
    people still need to know what they're buying though without having to get out a calculator and slide rule!

    i think most (non-uber-geeky) people would just go on the speed numerical (3000+, 3400+ etc) and the price and work out the rest from there depending on their budgets.

    Isn't that the point tho?

    At the very least they KNOW that a 3500 is faster than a 3200, and don't have to bother with knowing the difference between clock speed, cache, and single or dual channel is.

    Ever tried to explain what an FSB is, and why a faster one might be better, and none at all is probably best, to a non-techie?


  • Registered Users Posts: 4,003 ✭✭✭rsynnott


    astrofool wrote:
    Isn't that the point tho?

    At the very least they KNOW that a 3500 is faster than a 3200, and don't have to bother with knowing the difference between clock speed, cache, and single or dual channel is.

    Ever tried to explain what an FSB is, and why a faster one might be better, and none at all is probably best, to a non-techie?

    Well, they BELIEVE that 3500 is faster than 3200 ;)

    (Allegedly, the numbers are supposed to be based on comparisons to megahertz speeds on the old Athlons, or so AMD alleges; they are transparently a comparison to the P4 nowadays.)


  • Registered Users Posts: 16,608 ✭✭✭✭astrofool


    That was dropped when they left the Athlon XP, and it seemed to try and compare more directly with Intel, as an athlon 64 2800+ outperformed an Xp3200+.

    Given AMD's numbering method, a 3500+ will almost always be faster than a 3200+, the % difference will vary, but it will still be faster, a 3700+ will also be faster than it again.


  • Registered Users Posts: 4,003 ✭✭✭rsynnott


    They claim that that's just 'cause they stopped keeping it in step properly, I think. They're unwilling to admit that it's an Intel comparison. I wonder will it survive long into the multicore chips?


  • Moderators Posts: 5,555 ✭✭✭Azza


    Orginally the 2 3200's on socket 754 the 2Ghz (1mb Cache) and the 2.2Ghz (512kb) cache the one with more cache was dearer (slightly). Which seems to be reversed with the 3800 and 3700 on 939. Surely those 2 chips are equal in performance no?

    But why if the 3200 socket 754 was classified as the same chip (the 3200) why have they suddenly said hang on the chip with the higher clock speed is now faster and should have a faster rating (i.e 3800). Both the 3700 and 3800 are 90nm and have the same core tweaks and SSE3 support.


  • Advertisement
  • Registered Users Posts: 215 ✭✭Beno


    Now i'm more confused than before i read this thread :confused:


  • Registered Users Posts: 144 ✭✭r00t69


    Beno wrote:
    Now i'm more confused than before i read this thread :confused:

    Then my work here is done :D


Advertisement