comp.lang.idl-pvwave archive
Messages from Usenet group comp.lang.idl-pvwave, compiled by Paulo Penteado

Home » Public Forums » archive » QUESTION: is there a command in idl which could clean up memory pieces?
Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend 
Switch to threaded view of this topic Create a new topic Submit Reply
QUESTION: is there a command in idl which could clean up memory pieces? [message #62993] Tue, 21 October 2008 12:03 Go to next message
litongmu is currently offline  litongmu
Messages: 7
Registered: October 2008
Junior Member
Hi all,

I am wondering if there is a command in IDL which could clean up all
the memory pieces to generate a whole big piece which could be used by
IDL.
Because, recently I am running a code to read and deal with cdf files.
And there are lots of files (more than 1000). The code cannot finish
them once, because of memory problem.
And I know that in Matlab there is a command called 'pack', which
could clean up all the memory pieces. So I am wondering if there is a
similar counterpart in IDL. It took me really a long time to find in
the manuals. But no luck.
I appreciate your help and your try.

tongmu
Re: QUESTION: is there a command in idl which could clean up memory pieces? [message #63035 is a reply to message #62993] Thu, 23 October 2008 23:56 Go to previous message
Tatcher is currently offline  Tatcher
Messages: 7
Registered: September 2008
Junior Member
On Oct 24, 3:02 am, MC <Morefl...@gmail.com> wrote:
> Why not use plot and p.multi to reduce the number of windows and allow
> closure at will? How do you wade through so many plots at once anyway?
>
> Cheers
>
> On Oct 23, 9:09 pm, Tatcher <christian_ma...@gmx.at> wrote:
>
>
>
>
>
>> Hej!
>
>> I have a similar problem.
>
>> I am analyzing shots measured in a tokamak. The problem in my case is,
>> that the program opens up to 10 iplot windows per shot. Each window
>> contains 420000 data points. After approx. 15 shots, the Computer runs
>> out of memory (2 GB).
>> I have the same problem when I suppress the visualisation of the iplot
>> windows (user_interface='n') and dump the picture of the plot directly
>> to the harddisk instead.
>> Although my program runs in a loop and all the variables are defined
>> only once and are rewritten for each shot.In my case iPlot seems to be
>> the problem.Is it possible that every iPlot window reserve a certain
>> amount of memory? How do you plot your data?
>
>> My solution: I will ty to use float instead of double numbers and I
>> will upgraded my PC to 8 GB RAM.
>
>
Only small parts of a shot are of interest. My idea is to sight the
complete raw data in a first run and to choose the shots and time
intervals that look interesting. The advantage of iPlot is that I can
zoom in, annotate the plot and finally save it to the harddisk without
further coding. I think that is not possible with plot and p.multi in
an simple way.
In a second step I put promissing looking shotnumbers together with
the time ranges of interest, informations about the probes,
distances, ... in a spreadsheet and use this as a control file for my
program.
Another thing is that if I like to plot radial plasma potential
profiles for example that each shot gives just one point in the graph.
But in this case I can skip all other plots and have no memory
problem.

Maybe someone has another idea for a better and not so much memory
consuming concept.
Re: QUESTION: is there a command in idl which could clean up memory pieces? [message #63037 is a reply to message #62993] Thu, 23 October 2008 18:02 Go to previous message
MC is currently offline  MC
Messages: 50
Registered: September 1996
Member
Why not use plot and p.multi to reduce the number of windows and allow
closure at will? How do you wade through so many plots at once anyway?

Cheers

On Oct 23, 9:09 pm, Tatcher <christian_ma...@gmx.at> wrote:
> Hej!
>
> I have a similar problem.
>
> I am analyzing shots measured in a tokamak. The problem in my case is,
> that the program opens up to 10 iplot windows per shot. Each window
> contains 420000 data points. After approx. 15 shots, the Computer runs
> out of memory (2 GB).
> I have the same problem when I suppress the visualisation of the iplot
> windows (user_interface='n') and dump the picture of the plot directly
> to the harddisk instead.
> Although my program runs in a loop and all the variables are defined
> only once and are rewritten for each shot.In my case iPlot seems to be
> the problem.Is it possible that every iPlot window reserve a certain
> amount of memory? How do you plot your data?
>
> My solution: I will ty to use float instead of double numbers and I
> will upgraded my PC to 8 GB RAM.
Re: QUESTION: is there a command in idl which could clean up memory pieces? [message #63049 is a reply to message #62993] Thu, 23 October 2008 10:49 Go to previous message
pgrigis is currently offline  pgrigis
Messages: 436
Registered: September 2007
Senior Member
Tatcher wrote:
> On Oct 23, 3:48�pm, pgri...@gmail.com wrote:
>> Tatcher wrote:
>>> Hej!
>>
>>> I have a similar problem.
>>
>>> I am analyzing shots measured in a tokamak. The problem in my case is,
>>> that the program opens up to 10 iplot windows per shot. Each window
>>> contains 420000 data points. After approx. 15 shots, the Computer runs
>>> out of memory (2 GB).
>>> I have the same problem when I suppress the visualisation of the iplot
>>> windows (user_interface='n') and dump the picture of the plot directly
>>> to the harddisk instead.
>>> Although my program runs in a loop and all the variables are defined
>>> only once and are rewritten for each shot.In my case iPlot seems to be
>>> the problem.Is it possible that every iPlot window reserve a certain
>>> amount of memory? How do you plot your data?
>>
>> Why don't you close the iplot window after each plot?
>>
>> Paolo
>>
>>
>>
>>> My solution: I will ty to use float instead of double numbers and I
>>> will upgraded my PC to 8 GB RAm
>
> Unfortunately that is not possible. iPlot window are "locked" as long
> the program is executed. e.g. It's not possible to use the build in
> Zoom function, annotate the plot or simply close it.

Then you should make the program close it when it's done with it.

Paolo
Re: QUESTION: is there a command in idl which could clean up memory pieces? [message #63053 is a reply to message #62993] Thu, 23 October 2008 08:59 Go to previous message
Tatcher is currently offline  Tatcher
Messages: 7
Registered: September 2008
Junior Member
On Oct 23, 3:48 pm, pgri...@gmail.com wrote:
> Tatcher wrote:
>> Hej!
>
>> I have a similar problem.
>
>> I am analyzing shots measured in a tokamak. The problem in my case is,
>> that the program opens up to 10 iplot windows per shot. Each window
>> contains 420000 data points. After approx. 15 shots, the Computer runs
>> out of memory (2 GB).
>> I have the same problem when I suppress the visualisation of the iplot
>> windows (user_interface='n') and dump the picture of the plot directly
>> to the harddisk instead.
>> Although my program runs in a loop and all the variables are defined
>> only once and are rewritten for each shot.In my case iPlot seems to be
>> the problem.Is it possible that every iPlot window reserve a certain
>> amount of memory? How do you plot your data?
>
> Why don't you close the iplot window after each plot?
>
> Paolo
>
>
>
>> My solution: I will ty to use float instead of double numbers and I
>> will upgraded my PC to 8 GB RAm

Unfortunately that is not possible. iPlot window are "locked" as long
the program is executed. e.g. It's not possible to use the build in
Zoom function, annotate the plot or simply close it.
Re: QUESTION: is there a command in idl which could clean up memory pieces? [message #63058 is a reply to message #62993] Thu, 23 October 2008 06:48 Go to previous message
pgrigis is currently offline  pgrigis
Messages: 436
Registered: September 2007
Senior Member
Tatcher wrote:
> Hej!
>
> I have a similar problem.
>
> I am analyzing shots measured in a tokamak. The problem in my case is,
> that the program opens up to 10 iplot windows per shot. Each window
> contains 420000 data points. After approx. 15 shots, the Computer runs
> out of memory (2 GB).
> I have the same problem when I suppress the visualisation of the iplot
> windows (user_interface='n') and dump the picture of the plot directly
> to the harddisk instead.
> Although my program runs in a loop and all the variables are defined
> only once and are rewritten for each shot.In my case iPlot seems to be
> the problem.Is it possible that every iPlot window reserve a certain
> amount of memory? How do you plot your data?
Why don't you close the iplot window after each plot?

Paolo


>
> My solution: I will ty to use float instead of double numbers and I
> will upgraded my PC to 8 GB RAM.
Re: QUESTION: is there a command in idl which could clean up memory pieces? [message #63061 is a reply to message #62993] Thu, 23 October 2008 01:09 Go to previous message
Tatcher is currently offline  Tatcher
Messages: 7
Registered: September 2008
Junior Member
Hej!

I have a similar problem.

I am analyzing shots measured in a tokamak. The problem in my case is,
that the program opens up to 10 iplot windows per shot. Each window
contains 420000 data points. After approx. 15 shots, the Computer runs
out of memory (2 GB).
I have the same problem when I suppress the visualisation of the iplot
windows (user_interface='n') and dump the picture of the plot directly
to the harddisk instead.
Although my program runs in a loop and all the variables are defined
only once and are rewritten for each shot.In my case iPlot seems to be
the problem.Is it possible that every iPlot window reserve a certain
amount of memory? How do you plot your data?

My solution: I will ty to use float instead of double numbers and I
will upgraded my PC to 8 GB RAM.
  Switch to threaded view of this topic Create a new topic Submit Reply
Previous Topic: how to suppres warning
Next Topic: Re: PLOTS in different windows with different RANGE

-=] Back to Top [=-
[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Wed Oct 08 15:55:39 PDT 2025

Total time taken to generate the page: 0.00663 seconds