Of course, just as I sent my comment, I noticed you mentioned having
some symmetry in the problem. So a little more complete estimate of
memory requirements for a single precision executable on a machine
with 4 byte single precision is:
# Segs. Memory
20000 3062 MB
15000 1725 MB
10000 770 MB
5000 197 MB
2500 53 MB
---Paul
-----Original Message-----
From: Halsema, Paul
Sent: Tuesday, February 06, 2001 8:44 AM
Note that for such a large problem you will need about 3GB of memory
if you want to be able to solve it all "in core". An equivalent
amount of swap space and significantly longer run time can be used
instead.
---Paul
-----Original Message-----
From: John B. Wood [mailto:wood_at_itd.nrl.navy.mil]
Sent: Friday, February 02, 2001 10:20 AM
Hello, everyone. For about five years now I have run f77-compiled
NEC4 on an HP Apollo 755 workstation. I am looking to replace this
workstation with newer desktop hardware and would like to have the
capability to run a 20,000 (twenty thousand) segment (assuming some
symmetry) problem in less than 24 hours on equipment dedicated to
this task.
Received on Fri Feb 09 2001 - 14:28:28 EST
This archive was generated by hypermail 2.2.0 : Sat Oct 02 2010 - 00:10:41 EDT