Thanks to all who shared their experiences. Looks like the problem may not
arise with EVERY SINGLE application of matlab under solaris 2.5. My original
source indicated that "large" data arrays appeared to be the cause, or at
least a correlate. I am leaning towards proceeding with the upgrade, yet
keeping a careful eye on memory. Our user group on this system is
sufficiently small such that occasional reboots, if necessary, aren't a big
deal. Apparently version 5 will support solaris 2.5. Thanks again, Don
==================================================================
Donald R. Kobayashi, Fishery Biologist & P/T Systems Administrator
Fisheries Oceanography Investigation, NMFS/NOAA Honolulu Lab
2570 Dole Street, Honolulu HI 96822-2396
Voice:808-943-1242 FAX:808-943-1290 (dkobayas@mahi.nmfs.hawaii.edu)
Original inquiry:
We were planning to upgrade our sparc20 from solaris 2.4 to 2.5 until we
heard of a possible memory problem with matlab and solaris 2.5. Apparently,
memory does not release properly leading to some kind of slow, choking death.
Can anyone enlighten me as to their experiences with this? Perhaps some
clever soul has figured out a solution. Or perhaps the problem isn't
insidious enough to justify NOT upgrading. Comments anyone?
James Ashton <James.Ashton@keating.anu.edu.au> wrote:
I haven't seen this problem but I did hit a core dump using the
Symbolic Math toolbox. Referring this to the MathWorks revealed that
they do not support MATLAB 4.2 under Solaris 2.5. Nevertheless they
did eventually provide a fix for my problem. I'm unaware of any other
problems with the combination. Just be aware that the MathWorks will
not officially support Solaris 2.5 until MATLAB version 5 comes out.
That is likely to happen this calendar year but, last I heard, no
release date has been announced.
Sahir Siddiqui <sahirns@menger.eecs.stevens-tech.edu> wrote:
We haven't had any problems running Matlab on Sparc 20 and Sparc 5
machines with Solaris 2.5. We just upgraded a 690MP server from SunOS
4.1.3U1 to Solaris 2.5, and so far nobody has reported any Matlab
problems. I'd be interested in reading a summary of significant
responses you receive.
Dean Humphrey <dean@npg.wustl.edu> wrote:
I am transitioning from 4.1.3 to 2.5. A dozen or so are 2.5 now and the rest
are SunOS. We have several programs running on our departmental server
(SS10). So far I have not seen a problem. At the time I was upgrading and
installing vendor packages; folks at matlab did not yet support 2.5. In fact
I had the old version running and I then installed the 2.5 binaries and added
its license to the license manager. When I told matlab what I was planning
on doing; they asked me to call them back to let them know that it works!
Unitl all user workstations are upgraded; old OS machines automount the old
OS server disk and the new get the 2.5 server disk (usr/local etc.) So far so
good! p.s. This has been working for 5 months now.
GREGOR E FELLERS <u2is9gef@gregsun.crrel.usace.army.mil> wrote:
We have matlab running on a SPARC 20 (two processors)
and 2.5. I have not had one complaint as you
describe.
"Dr.W.Tych" <w.tych@lancaster.ac.uk> wrote:
I must have missed something here, but (maybe because I missed it) Matlab
seems to work quite happily on two Solaris 2.5 machines I know.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:02 CDT