lkml.org 
[lkml]   [1997]   [Jul]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: Kernel cpu selection + other platforms?
    On Mon, 21 Jul 1997, Roland Steinbach wrote:

    > Mike Jagdis wrote:
    >
    > > Why? It is pretty easy to recognise all x86 varieties (other than the
    > > early 486 clones) at run time and simply Do The Right Thing without
    > > asking the user daft questions about whether they have microcode
    > > bug #564 or working frozzle optimizations.

    It is very easy to identify the CPU at run time but you can't do anything
    about what you find. It's too late. To obtain optimum performance, you
    need the code compiled to take advantage of a particular CPU. For
    instance, a '486 doesn't give a damn about longword alignment. A pentium
    does. You can't realign code at run time. It is possible to align buffers
    at run time, but the overhead could be somewhat extreme. You can't just
    build a kernel for the latest and greatest CPU because some internally-
    locked instructions, which you really would like to use, are not supported
    in a '386, etc., etc.

    So with Intel, we have to make a kernel which will run with anything, then
    rebuild the sources to take advantage of the platform's latest CPU
    advances.


    Cheers,
    DJ
    -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
    Richard B. Johnson
    Analogic Corporation
    Email : rjohnson@analogic.com, johnson@analogic.com
    Penguin : Linux version 2.1.44 on an i586 machine (66.15 BogoMips).
    Warning : It's hard to stay on the trailing edge of technology.
    -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


    \
     
     \ /
      Last update: 2005-03-22 13:39    [W:2.558 / U:0.012 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site