Re: Splus on a Sun via Solaris 2.6

Steve Boswell (boswell@ll.mit.edu)
Tue, 20 Jan 98 10:42:03 -0500


At 8:37 AM -0600 1/20/98, <HumboltA@kochind.com> wrote:
>My questions today are specific to those of you who are running Splus
>using newer Solaris operating systems.
>
>We've installed Splus on a new Sun Ultra Enterprise 450 server using the
>Solaris 2.6 operating system. I have run a large in-house function (no
>C or Fortan compiled programs and all basic Splus code) which initially
>takes 153 seconds on this new Sun. If I continue to run exactly the
>same function on exactly the same data, however, it continues to require
>longer and longer times. E.g., by the 6th execution it takes 485
>seconds to run. This is the basis for my concern and questions. [ The
>function works fine on an old IBM RS/6000. It also works fine on Splus
>V4.0, which actually runs the function in just 130 seconds on my 200 MHz
>PC. ]
>
>Has anyone had similar experiences on the Solaris 2.6 OS? Are we
>advised to retreat to an earlier OS? Any hints as to what we might try
>next?
>
>On another note, does anyone know where we can get an emacs editor for
>Solaris 2.6? How about for Solaris 2.5.1 if we retreat to that OS? Or,
>can we easily use the Solaris textedit editor from within Splus?
>
>Allen Humbolt
>Quantitative Analyst
>Koch Industries, Inc.
>HumboltA@kochind.com e-mail
>(316) 828-7911 phone
>(316) 828-8674 fax

Allen,

You can get emacs from the Free Software Foundation site at MIT,
prep.ai.mit.edu,
using anonymous FTP. Emacs distributions reside in the
directory pub/gnu. You can download a late model, download it, and
install it on your machine. I wouldn't think that Solaris 2.6 vs 2.5
would be an issue. There is also an introductory file.
GETTING.GNU.SOFTWARE, in that directory.

Hope this helps.

Steve Boswell
Lincoln Laboratory, MIT
boswell@ll.mit.edu

BTW, I am curious as to what a quantitative analyst at Koch Industries
does. I am interested in financial engineering/pricing. Thanks.