| Re: Memory Leakage......(I think) [message #30477] |
Wed, 01 May 2002 09:11 |
Rick Towler
Messages: 821
Registered: August 1998
|
Senior Member |
|
|
Have you tried removing the calls to the temporary function?
-Rick
"Andrew Fielding" <a.fielding@icr.ac.uk> wrote in message
news:3d9ed221.0205010403.406fe470@posting.google.com...
> Thanks for the comments regarding my problem. Someone suggested I post
> the code to clarify the problem so here it is.
> This is a loop within a sub-procedure that is called (256*256) times
> from the main procedure. The count variable controlling the loop
> varies from a minimum of 0 up to 510 depending on the call. I am
> fairly certain the problem is with the arithmetic within the loop as
> removing it and replacing it with a simple statement results in the
> main program processing in around 2 minutes! Any hints/suggestions
> gratefully received.
>
> Andrew
>
> ; Weighted Interpolation voxel density calculation
>
> fval=0.0
> sum=0.0
>
> for i=0L,count-1 do begin
> index=ishft(X1,-14) + row[ishft(Y1,-14)] + slice[ishft(Z1,-14)]
> val=ctcube[index]
> if (val gt config.air) then begin
> d0= X1/SCALE - fix(X1/SCALE)
> d1= Y1/SCALE - fix(Y1/SCALE)
> d2= Z1/SCALE - fix(Z1/SCALE)
> fval=(1.0-d0)*(1.0-d1)*(1.0-d2)*float(val)
> fval=temporary(fval) + d0*(1.0-d1)*(1.0-d2)*float(ctcube[index+1])
> fval=temporary(fval) +
> (1.0-d0)*d1*(1.0-d2)*float(ctcube[index+xdim])
> fval=temporary(fval) + d0*d1*(1.0-d2)*float(ctcube[index+1+xdim])
> index=temporary(index) + xdim*ydim
> fval=temporary(fval) + (1.0-d0)*(1.0-d1)*d2*float(ctcube[index])
> fval=temporary(fval) + d0*(1.0-d1)*d2*float(ctcube[index+1])
> fval=temporary(fval) + (1.0-d0)*d1*d2*float(ctcube[index+xdim])
> fval=temporary(fval) + d0*d1*d2*float(ctcube[index+1+xdim])
>
> val= fix(fval)
> sum=temporary(sum) + convert_ct2mu[val-(Hmin-1024)]
> endif
>
> X1=X1+dX
> Y1=Y1+dY
> Z1=Z1+dZ
> endfor
|
|
|
|
| Re: Memory Leakage......(I think) [message #30479 is a reply to message #30477] |
Wed, 01 May 2002 05:03  |
a.fielding
Messages: 4
Registered: May 2002
|
Junior Member |
|
|
Thanks for the comments regarding my problem. Someone suggested I post
the code to clarify the problem so here it is.
This is a loop within a sub-procedure that is called (256*256) times
from the main procedure. The count variable controlling the loop
varies from a minimum of 0 up to 510 depending on the call. I am
fairly certain the problem is with the arithmetic within the loop as
removing it and replacing it with a simple statement results in the
main program processing in around 2 minutes! Any hints/suggestions
gratefully received.
Andrew
; Weighted Interpolation voxel density calculation
fval=0.0
sum=0.0
for i=0L,count-1 do begin
index=ishft(X1,-14) + row[ishft(Y1,-14)] + slice[ishft(Z1,-14)]
val=ctcube[index]
if (val gt config.air) then begin
d0= X1/SCALE - fix(X1/SCALE)
d1= Y1/SCALE - fix(Y1/SCALE)
d2= Z1/SCALE - fix(Z1/SCALE)
fval=(1.0-d0)*(1.0-d1)*(1.0-d2)*float(val)
fval=temporary(fval) + d0*(1.0-d1)*(1.0-d2)*float(ctcube[index+1])
fval=temporary(fval) +
(1.0-d0)*d1*(1.0-d2)*float(ctcube[index+xdim])
fval=temporary(fval) + d0*d1*(1.0-d2)*float(ctcube[index+1+xdim])
index=temporary(index) + xdim*ydim
fval=temporary(fval) + (1.0-d0)*(1.0-d1)*d2*float(ctcube[index])
fval=temporary(fval) + d0*(1.0-d1)*d2*float(ctcube[index+1])
fval=temporary(fval) + (1.0-d0)*d1*d2*float(ctcube[index+xdim])
fval=temporary(fval) + d0*d1*d2*float(ctcube[index+1+xdim])
val= fix(fval)
sum=temporary(sum) + convert_ct2mu[val-(Hmin-1024)]
endif
X1=X1+dX
Y1=Y1+dY
Z1=Z1+dZ
endfor
|
|
|
|
| Re: Memory Leakage......(I think) [message #30485 is a reply to message #30479] |
Tue, 30 April 2002 16:42 |
Peter Scarth
Messages: 9
Registered: February 2000
|
Junior Member |
|
|
Hi Andrew,
I have experienced a similar problem a while back with some code involving a
large number of loops processing a large image. IDL would exit after a
couple of hours on my win2k system. I never worked out exactly why it would
just stop (I never wanted to wait 4 hours for it to see if it was memory or
resources related), but removing the following two lines from my code fixed
it:
caldat,st+(systime(1,/julian)-st)*100./(50.*((n-1.)*r+i)/r), mon,day,yr,hr,mi
,sec
progressBar->SetLabel, "Completion Time>
"+strtrim(string(hr),2)+":"+strtrim(string(mi),2)
This just put up an estimate of when the code would complete. I *know* the
problem couldn't be in David's progress bar. I'm pretty sure the problem was
in the systime call - I was calling it way too often. I suspect this is a
windows problem, but I don't know...
I also had heaps of problems in win95 a few years back when I was using a
Logitech scroll mouse. After about 15 mins of scrolling in IDL I'd run out
of system resources. Weird.
Cheers,
Peter.
"Andrew Fielding" <a.fielding@icr.ac.uk> wrote in message
news:3d9ed221.0204300344.3be3a759@posting.google.com...
> I am running IDL 5.5 under Windows 95 and have following problem. I
> have a loop that is repeated a large number of times and the loop
> contains a series of floating point multiplications. The problem is
> that the pc grinds slowly to a halt as the number of iterations
> increases, eventualy, stopping completely. Is there a known operating
> system issue with not releasing memory or is this a programming error
> on my part.
|
|
|
|
| Re: Memory Leakage......(I think) [message #30488 is a reply to message #30485] |
Tue, 30 April 2002 14:32 |
Mark Hadfield
Messages: 783
Registered: May 1995
|
Senior Member |
|
|
"David Fanning" <david@dfanning.com> wrote in message
news:MPG.17383bc056468d09898b2@news.frii.com...
> Andrew Fielding (a.fielding@icr.ac.uk) writes:
>
>> I am running IDL 5.5 under Windows 95 and have following
>> problem. ...
>
> It's a programming error. Oh, wait... Windows 95!? It's *probably*
> a programming error. :-)
I vote for the programming error too. Andrew, can you please tell us
more about what you're doing. Post the code!
--
Mark Hadfield "Ka puwaha et tai nei, Hoea tatou"
m.hadfield@niwa.co.nz
National Institute for Water and Atmospheric Research (NIWA)
|
|
|
|
| Re: Memory Leakage......(I think) [message #30494 is a reply to message #30488] |
Tue, 30 April 2002 05:28 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Andrew Fielding (a.fielding@icr.ac.uk) writes:
> I am running IDL 5.5 under Windows 95 and have following problem. I
> have a loop that is repeated a large number of times and the loop
> contains a series of floating point multiplications. The problem is
> that the pc grinds slowly to a halt as the number of iterations
> increases, eventualy, stopping completely. Is there a known operating
> system issue with not releasing memory or is this a programming error
> on my part.
It's a programming error. Oh, wait... Windows 95!?
It's *probably* a programming error. :-)
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
