| Re: Efficient Programming [message #4615] |
Sun, 02 July 1995 00:00 |
Leonard Kramer
Messages: 2
Registered: July 1995
|
Junior Member |
|
|
dean@phobos.cira.colostate.edu wrote:
>
> What would be the most efficient way to get this to work in IDL. This is
> obvious slow with the FOR DO BEGIN.....
>
>
> Kelly Dean
> CSU/CIRA
>
> ============================================================ ===============
>
> nx = 1024
> ny = 768
> North_bound = 3429
> West_bound = 9249
> line_res = 4
> elem_res = 8
> LINarr = LONarr(nx*ny)
> PIXarr = LONarr(nx*ny)
> k = 0L
> * FOR y = 0, ny-1 DO BEGIN
> * FOR x = 0, nx-1 DO BEGIN
> * LINarr(k) = North_bound + ( y * line_res)
> * PIXarr(k) = West_bound + ( x * elem_res)
> * k = k + 1L
> * ENDFOR
> * ENDFOR
>
You see those lines I marked with an * above. Replace them
with:
|
|
|
|
| Re: Efficient Programming [message #4616 is a reply to message #4615] |
Sun, 02 July 1995 00:00  |
Leonard Kramer
Messages: 2
Registered: July 1995
|
Junior Member |
|
|
dean@phobos.cira.colostate.edu wrote:
>
> What would be the most efficient way to get this to work in IDL. This is
> obvious slow with the FOR DO BEGIN.....
>
>
> Kelly Dean
> CSU/CIRA
>
> ============================================================ ===============
>
> nx = 1024
> ny = 768
> North_bound = 3429
> West_bound = 9249
> line_res = 4
> elem_res = 8
> LINarr = LONarr(nx*ny)
> PIXarr = LONarr(nx*ny)
> k = 0L
> * FOR y = 0, ny-1 DO BEGIN
> * FOR x = 0, nx-1 DO BEGIN
> * LINarr(k) = North_bound + ( y * line_res)
> * PIXarr(k) = West_bound + ( x * elem_res)
> * k = k + 1L
> * ENDFOR
> * ENDFOR
>
You see those lines I marked with an * above. Replace them
with:
|
|
|
|
| Re: Efficient Programming [message #4617 is a reply to message #4615] |
Sun, 02 July 1995 00:00 |
hebeling
Messages: 1
Registered: July 1995
|
Junior Member |
|
|
In article <3t5b5h$2ts4@yuma.ACNS.ColoState.EDU>, dean@phobos.cira.colostate.edu writes:
>
> What would be the most efficient way to get this to work in IDL. This is
> obvious slow with the FOR DO BEGIN.....
>
>
> Kelly Dean
> CSU/CIRA
>
> ============================================================ ===============
>
> nx = 1024
> ny = 768
> North_bound = 3429
> West_bound = 9249
> line_res = 4
> elem_res = 8
> LINarr = LONarr(nx*ny)
> PIXarr = LONarr(nx*ny)
> k = 0L
> FOR y = 0, ny-1 DO BEGIN
> FOR x = 0, nx-1 DO BEGIN
> LINarr(k) = North_bound + ( y * line_res)
> PIXarr(k) = West_bound + ( x * elem_res)
> k = k + 1L
> ENDFOR
> ENDFOR
>
Try this:
nx = 1024L ;ought to be LONGWORD
ny = 768L ;ought to be LONGWORD
North_bound = 3429
West_bound = 9249
line_res = 4
elem_res = 8
linarr = replicate(1,nx) # (lindgen(ny)*line_res) + north_bound
pixarr = (lindgen(nx)*elem_res) # replicate(1,ny) + west_bound
linarr = reform(linarr,nx*ny) ;force array to be 1-dim
pixarr = reform(pixarr,nx*ny) ;force array to be 1-dim
That's at least an order of magnitude faster than the double FOR loop.
Harald Ebeling
=======================
University of Cambridge
Institute of Astronomy
Madingley Road
Cambridge CB3 0HA
United Kingdom
ebeling@ast.cam.ac.uk (Ignore the address in the header - I'm only visiting...)
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
