Hand coding of routines
Christopher Browne
cbbrowne-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org
Wed Nov 30 02:03:24 UTC 2005
On 11/29/05, Ian Zimmerman <nobrowser-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org> wrote:
>
> Peter> 3. Yet another person whose previous job was in the patent office
> Peter> (sounds familiar?)
>
> Lennart> Can't think of anything here. Nope nothing familiar pops to
> Lennart> mind.
>
> He probably means Einstein. But sorry, there's no comparison.
>
> To me, Mr. Goto's effort seems like a waste of life. But as long as it
> makes _him_ happy, more power to him.
Historically, a remarkable amount of human work has gone into
optimizers for the FORTRAN decks used to solve big problems. People
have been working on NETLIB code and such for decades. There isn't
much change now, but a LOT of effort went into the FORTRAN compilers
that process these things.
His time may in retrospect appear wasted in view of the way Digital
squandered their technical heritage. Alpha was at the point where it
was plenty mature enough to warrant such tuning efforts were it to
survive, commercially.
You and I may not use NETLIB-oriented code terribly much; that doesn't
mean it doesn't have a significant set of users that use it a lot who
would benefit from some pretty low level tuning.
Mind you, a fair number of those users at relevant supercomputing labs
like Fermi and Sandia are doing nuclear device design; that might not
be high on everyone's preferences :-).
--
http://www3.sympatico.ca/cbbrowne/linux.html
"The true measure of a man is how he treats someone who can do him
absolutely no good." -- Samuel Johnson, lexicographer (1709-1784)
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
More information about the Legacy
mailing list