| IDL memory usage [message #9080] |
Wed, 28 May 1997 00:00  |
Gary Fu
Messages: 9
Registered: April 1997
|
Junior Member |
|
|
Hi,
I monitored the memory usage on the following two IDL cases.
In case 1, the IDL process takes 31MB of memory (8MB of IDL
reserved + 24MB of data, reasonable). However, in case 2,
it takes 47MB of memory. Anyone knows why and is there a way
to fix the problem ?
Thanks.
Gary
**********************************************************
Case 1:
a = indgen(2000,4000)
a = a(*,0:1999)
Case 2:
a = indgen(2000,4000)
ia = indgen(2000) * 2
b = a(*,ia) ; OR b = TEMPORARY(a(*,ia))
**********************************************************
--
*****************************************
* Gary Fu, GSC (301) 286-7107 *
* email : "gfu@seadas.gsfc.nasa.gov" *
* NASA/Goddard Space Flight Center *
*****************************************
|
|
|
|
| Re: IDL memory usage [message #37632 is a reply to message #9080] |
Wed, 14 January 2004 07:07  |
R.G. Stockwell
Messages: 363
Registered: July 1999
|
Senior Member |
|
|
"Michael Wallace" <mwallace.removethismunge@swri.edu.invalid> wrote in message news:bu1m97$7jd$1@nntp1.jpl.nasa.gov...
...
> I have a process which uses a fairly constant amount of
> memory (in IDL), but IDL keeps claiming more and more memory from the OS
> as time goes on. ...
IMHO the preferred solution would be to find the
programming error that is causing the memory leak.
I run IDL for weeks and weeks with no problem.
Check your heap, do you have pointers or objects
that are not being cleaned up?
Cheers,
bob
|
|
|
|
| Re: IDL memory usage [message #37638 is a reply to message #9080] |
Tue, 13 January 2004 17:04 |
Michael Wallace
Messages: 409
Registered: December 2003
|
Senior Member |
|
|
> However, this may not have the desired behavior. Once the program
> reaches the limit, any further memory allocations will fail and may
> wreak havoc on your program. It won't just run "slower."
In this particular case, I don't believe anything will be creating
havoc. I did a little test. First, I ran IDL without much else going
on. Eventually, it filled up most of the available memory. Once
finished, I opened up a bunch of other programs which chewed up a good
chunk of memory. I then ran my same IDL program again. It again filled
up most of the available memory, but it had nowhere near the amount that
it had the first time around (since other programs had already claimed
some of that) and it still executed in about the same time. So, this
led me to conclude that I might be able to safely limit the IDL process
and not run into any problems. Also, I was periodically printing out
the IDL memory usage from within IDL and it leveled off early on and
never grew over the rest of the execution even though you could watch
more and more memory being taken from the OS during this same time.
I will try limit/ulimit and see what happens. Thanks.
Mike
|
|
|
|
| Re: IDL memory usage [message #37640 is a reply to message #9080] |
Tue, 13 January 2004 15:24 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
Michael Wallace <mwallace.removethismunge@swri.edu.invalid> writes:
> Is there any way to set a maximum limit on the amount of memory an IDL
> process will use even if the machine running the code has more memory
> available? I have a process which uses a fairly constant amount of
> memory (in IDL), but IDL keeps claiming more and more memory from the OS
> as time goes on. I'd just like to cap the amount of memory IDL can get
> its grubby little hands on even if it means that my code will execute a
> little slower.
If you are running on Unix, you can limit the memory consumption by
using the limit or ulimit commands, depending on the shell you are
using. You would run one of these commands in the shell before
running IDL.
However, this may not have the desired behavior. Once the program
reaches the limit, any further memory allocations will fail and may
wreak havoc on your program. It won't just run "slower."
Good luck,
Craig
--
------------------------------------------------------------ --------------
Craig B. Markwardt, Ph.D. EMAIL: craigmnet@REMOVEcow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
------------------------------------------------------------ --------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
