[desktop-discuss] Re: SuSE Enterprise Desktop 10 reviews

Stephen Hahn sch at eng.sun.com
Thu Jul 13 12:06:17 PDT 2006


* Bart Smaalders <bart.smaalders at Sun.COM> [2006-07-13 09:08]:
> Gerald Henriksen wrote:
> >On Thu, 13 Jul 2006 15:41:00 +0100, you wrote:
> >
> >>On 7/13/06, Gerald Henriksen <ghenriks at gmail.com> wrote:
> >>>
> >>>Despite what some people believe compiling code for a particular
> >>>version of the x86 cpu line doesn't offer a noticable improvement.
> >>Well that's not entirely true - compiling for amd64 can show a very
> >>significant boost. Which raises the question - why don't we compile
> >
> >But amd64/x64 is not the same as x86
> >
> >Which is why Linux and Windows have both gone to a separate product
> >for the x64 processor because the changes are significant enough to
> >justify/require native x64 binaries.
> >
> 
> We support both ISAs transparently on the same platform, just like
> we did for 64 bit sparc.
> 
> How many apps get 64 bit versions is a matter of resources.

  ... is a matter of getting contributors' attention and interest, I
  would say.

  Specifically, there are two ways to contribute.  One is identification
  of a potential problem:  if you believe a particular application
  should have a 64-bit version for performance or functionality reasons,
  file a bug report saying so.  The second is assessment of that
  potential problem:  if you see such a bug report and have make(1) and
  C language experience, consider making the (usually few) changes and
  verifying whether the application is improved, slightly or
  substantially.  If so, get the changes integrated; if not, update the
  bug with "not worth it".

  (sort(1) is an example of an application that was converted to 64-bit
  for functionality reasons.)

  - Stephen

-- 
Stephen Hahn, PhD  Solaris Kernel Development, Sun Microsystems
stephen.hahn at sun.com  http://blogs.sun.com/sch/



More information about the desktop-discuss mailing list