Loading ...
Sorry, an error occurred while loading the content.

Re: [TwinCLinG] Commercial Use of Linux in India

Expand Messages
  • Hrishikesh Das
    infact 256kb at full speed is faster than 512kb at half speed, this is because of other features of the coppermine, like different associativity mapping of the
    Message 1 of 21 , Nov 6, 2000
    • 0 Attachment
      infact 256kb at full speed is faster than 512kb at half speed, this is
      because of other features of the coppermine, like different associativity
      mapping of the cache and improved branch prediction,

      infact it was due to these that Intel p3 was able to cross the performance
      of athlons in 3d games,

      Hrishi


      ----- Original Message -----
      From: N. Pari Purna Chand <paripurnachand@...>
      To: <ilughyd@egroups.com>
      Sent: Monday, November 06, 2000 6:17 AM
      Subject: Re: [TwinCLinG] Commercial Use of Linux in India


      >
      > On Sun, 5 Nov 2000, Hrishikesh Das wrote:
      > > fcpga is for the Intel processors, OK not just for P3's, its for
      celerons
      > > and p3's
      > >
      > > but the fact that 256kb cache is not exclusive to fcpga only, it is
      > > characteristic of the coppermine processor and is not limited to fcpga
      only,
      > > Cu processors of slot1 also came with 256kb cache,
      > > ..............
      > > ......
      >
      > Intel did reduce the cache from 512 k to 256 k but
      > what the Inc says is that when it was using 512 kb,
      > the cache used to work only at half the speed of processor
      > and now this 256K works at full speed, resulting in the
      > same kinda performance.
      >
      > Chandu
      > --------------------------------------------------------------
      > L I N U X | N. Pari Purna Chand
      > The Planetary Operating System | paripurnachand@...
      >
      > In most countries selling harmful things like drugs is
      > punishable, then how come people can sell M$ software
      > and go unpunished ?
      > --------------------------------------------------------------
      >
      >
      >
      > ---------------------------------------------------
      > Check out http://linuxindia.virtualave.net
      >
      > If U want to Unsubscribe from the list please
      > send a mail to ilughyd-unsubscribe@egroups.com.
      > If U still find a problem please send a message to
      > ilughyd-owner@egroups.com
      >
      > Check Out our new irc channel
      > irc.slashnet.org #twincling (Happy Chatting)
      >
      >
    • Nick Hill
      hunh! flame time again eh?!?! ; ... if u know any bit of english, i said reduced . when i say reduced down to 256k , it means that something which has a
      Message 2 of 21 , Nov 6, 2000
      • 0 Attachment
        hunh! flame time again eh?!?! ;>

        On Sun, Nov 05, 2000 at 11:44:50PM +0530, Hrishikesh Das wrote:
        > <flame>
        >
        > ok now u mention intel reduced the l2 cache down to 256k, this is true for
        > P3's only 'cause celerons have 128kb, Now did U mention this ? for p3's or
        > for celerons's ?

        if u know any bit of english, i said "reduced". when i say "reduced down to
        256k", it means that something which has a higher value (p3s in this case) has
        been "reduced" to a lower value (256k in this case).

        Bah. ;P

        Nikhil.
      • Nick Hill
        This is going waaaaaaaay OT. if u want to continue this thread, take it offline ... i doubt the cache will make much of a difference when the geometric
        Message 3 of 21 , Nov 6, 2000
        • 0 Attachment
          This is going waaaaaaaay OT. if u want to continue this thread, take it offline

          On Mon, Nov 06, 2000 at 02:12:00PM +0530, Hrishikesh Das wrote:
          > infact it was due to these that Intel p3 was able to cross the performance
          > of athlons in 3d games

          i doubt the cache will make much of a difference when the geometric calculations
          are downloaded off to the accelerator (like the Radeon), as _most_ of the ops.
          are _not_ performed by the processor. An increase in the accelerator's cache
          sure matters a lot in this case.

          Nikhil.
        Your message has been successfully submitted and would be delivered to recipients shortly.