In article <(E-Mail Removed)>, "geoff" <(E-Mail Removed)> wrote:
>Sam Adams of IBM, here:
>
>http://www.stic.st/conferences/stic1...rallelism-obje
>ct-oriented-programming/
>
>. . . talks about how the physical barriers have been reached for computing
>and either new technology is needed or a new way to use current technology,
>such as data-centric computing (rather than cpu centric).

Maybe the limits will prompt efforts to program more efficiently?
On daily life computing tasks, bloatware kept up with the increases
in processing power to keep overall speeds approximately constant.
Which means that the next easiet thing is to make the software run
faster. There is probably a lot of room in there.

DK

On Apr 25, 2:10 am, d...@no.email.thankstospam.net (DK) wrote:
> In article <OqCdnZG17IIFMgvSnZ2dnUVZ_vOdn...@earthlink.com>, "geoff" <nos...@nospam.com> wrote:
> >Sam Adams of IBM, here:
>
> >http://www.stic.st/conferences/stic1...massive-parall...
> >ct-oriented-programming/
>
> >. . . talks about how the physical barriers have been reached for computing
> >and either new technology is needed or a new way to use current technology,
> >such as data-centric computing (rather than cpu centric).
>
> Maybe the limits will prompt efforts to program more efficiently?
> On daily life computing tasks, bloatware kept up with the increases
> in processing power to keep overall speeds approximately constant.
> Which means that the next easiet thing is to make the software run
> faster. There is probably a lot of room in there.
>
> DK

No there isn't. Predictive processing as a distributional aspect of
multi-core assignment has proven notoriously difficult for code
writers to otherwise gain substantial efficiency.

On Apr 25, 4:10*pm, d...@no.email.thankstospam.net (DK) wrote:
> In article <OqCdnZG17IIFMgvSnZ2dnUVZ_vOdn...@earthlink.com>, "geoff" <nos....@nospam.com> wrote:
> >Sam Adams of IBM, here:
>
> >http://www.stic.st/conferences/stic1...massive-parall...
> >ct-oriented-programming/
>
> >. . . talks about how the physical barriers have been reached for computing
> >and either new technology is needed or a new way to use current technology,
> >such as data-centric computing (rather than cpu centric).
>
> Maybe the limits will prompt efforts to program more efficiently?
> On daily life computing tasks, bloatware kept up with the increases
> in processing power to keep overall speeds approximately constant.
> Which means that the next easiet thing is to make the software run
> faster. There is probably a lot of room in there.
>
> DK


Yes, imagine if things were done in assembler and running on a 4Ghz
processor found in your typical PC.. Only problem is that it might
take years to release a particular software.