Was Moore's law, now something else, parallelism

[Jerry Feldman]

On 07/13/2010 10:26 AM, Dan Ritter wrote:
>> > A number of processes make sense for parallel processing, but its ha=
rd=20
>> > to do. People in the industry are complaining that it is a "language=
=20
>> > issue" in that we don't have the languages to express parallel=20
>> > processing well. Maybe.
>>    =20
> It's a brain issue, because it's hard for users and programmers to thin=
k
> in parallel terms.
>
>  =20
We've had language extensions for parallel computing for many years,
such as Posix Threads as well as spawning and forking. We've had
compilers, assemblers, and chips that can reorder instructions to
improve parallel computing.

In my experience, most application programmers have little or no
experience with threads, or other forms of parallelism. In addition, a
lot of the stock math libraries, like LAPACK, don't take parallelism
into account. But, in support of Mark's comment, a good parallelized
application needs to take parallelism into account by design. Most
modern computer languages are relatively easy to program for threads or
forking, but a poor design could make the entire parallelized
applicaiton slower.

--=20
Jerry Feldman <gaf-mNDKBlG2WHs at public.gmane.org>
Boston Linux and Unix
PGP key id: 537C5846
PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB  CA3B 4607 4319 537C 5846