[Date Index] [Thread Index] [Author Index]

Re: Hardware Mma ?

• To: mathgroup at yoda.physics.unc.edu
• Subject: Re: Hardware Mma ?
• From: jmg at cs.ucla.edu (James M Goodwin)
• Date: Tue, 26 May 92 15:46:29 -0700

|Perhaps, we are encouraged to use the fastest CPUs, like HP 9000/700 or
|forthcoming 80586, if we need much more speed-up.

It seems to me that Mma needs to have a compiler available to
use on finished code for production runs, and to provide compiled
rather than interpreted plugin units.  Is such a thing in the works?
Without that Mma is only useful for preliminary exploration in
many problems.  It seems unreasonable to have to recode in C or LISP
to get the speed needed to explore a problem, and it requires
reinventing all of the algorithms already available in Mma.

Any prospects from "Wolfram"?

-- jmg