Re: AMD64-based laptops from HP/Compaq are crippled

From: Russell Shaw (rjshawN_o_at_s_pam.netspace.net.au)
Date: 07/06/05

  • Next message: lobotomy: "Re: toshiba satellite ethernet linux"
    Date: Wed, 06 Jul 2005 16:41:49 +1000
    
    

    Wes Newell wrote:
    > On Tue, 05 Jul 2005 13:27:03 -0700, alex goldman wrote:
    >
    >>Wes Newell wrote:
    >>
    >>>Overall performance
    >>>increase for dual over single channel is 5% at most
    >>
    >>Where did you get this idea?
    >
    > From my considerable knowledge of how they work. And then there are
    > thounsands of benchmarks to prove it all over the web. find a benchmark of
    > apps that have a 3400+ single channel cpu and a 3500+ dual channel listed
    > in the same benchmark. Now guess which one wins most of the benchmark
    > test. if you think the 3500+ you'd be wrong. Now compare the 3400+ to a
    > 3800+. Both have the same clockspeed of 2.4GHz, and the same 512K cache.
    > Only difference is the dual channel ram. Does the 3800+ win all
    > benchmarks? No. And the reason is pretty simple. The overhead of the dual
    > channel slows it down when high bandwidth isn't needed. And the onboard
    > very low latency memory controller allows the single channel to service
    > the data request very fast compared to old chipset controller that had to
    > send the data over a shared FSB. The lack of a major performance increase
    > for dual channel is probably why HP decided not to use it if that's what
    > they did. I still think leaving it out was stupid though. But in reality,
    > it's not a big deal or advantage. Dual cores may prove to be a different
    > story since they will have much more demand. Time will tell.

    It may lead to a greater increase in battery lifetime compared to
    speed lost.


  • Next message: lobotomy: "Re: toshiba satellite ethernet linux"

    Relevant Pages

    • Re: Another Memory question!
      ... If you want a "pure" dual channel mode, ... part of the memory space is currently being used by your ... Designing a benchmark to pinpoint the differences would ... where the applications are storing their data in memory, ...
      (alt.comp.periphs.mainboard.asus)
    • Re: Memory...single channel DDR2 VS dual channel DDR2
      ... Is it worth my while performance wise to swap in a dual channel ... has no IDE for my optical drive and I would have to get a SATA CD/DVD ... the opportunity to do some benchmark testing. ... dual and single channel on the benchmark was only about 10%, ...
      (microsoft.public.windowsxp.hardware)