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

Home » Public Forums » archive » Re: Threads in IDL 7.0
Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend 
Return to the default flat view Create a new topic Submit Reply
Re: Threads in IDL 7.0 [message #63102 is a reply to message #63091] Tue, 28 October 2008 01:28 Go to previous messageGo to previous message
Bernhard Reinhardt is currently offline  Bernhard Reinhardt
Messages: 26
Registered: October 2008
Junior Member
Heinz Stege wrote:
> On Mon, 27 Oct 2008 16:44:29 +0100, Bernhard Reinhardt wrote:
>
>> Well, I wasn�t really precise. I�m not doing this on a 1d-array but on a
>> 4d-array, where 2 dimensions are time and 2 dimensions are space. I try
>> to filter special events in time and count those on a 2d-map. Here�s the
>> code:
>>
>> for i = 0, N_ELEMENTS(STRUC.data[*,0,0,0])-1 do begin
>> for j = 0, N_ELEMENTS(STRUC.data[0,*,0,0])-1 do begin
>> indices = Where(STRUC.data[i,j,*,*] GE 150., count)
>> freq [i,j]=count
>> endfor
>> endfor
>>
>> Although the array data is quite big "where" only get�s a small portion
>> to see of it. So thread-pool isn�t invoked. => CPU-Usage still 50%
>>
>> I also asked some more IDL-experienced colleagues about generating
>> threads manually but they also didn�t know about anything like that :(
>>
>> BUT using your method still brought me a gain of 3.6 times faster
>> execution :)
>>
>> regards
>>
>> Bernhard
>
> Please try the following command:
>
> freq=total(total(STRUC.data ge 150.,4,/integer),3,/integer)
>
> The IDL manual says, that TOTAL makes use of IDL�s thread pool. And
> there is no for-loop needed anymore...

Now that�s fast! Who cares about threads? ;) Thanks to everyone.

Regards

Bernhard
[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Re: Typeconversion within structure?
Next Topic: Re: DEPTH_TEST_FUNCTION, DEPTH_TEST_DISABLE properties (send an object to back)

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

Current Time: Sun Nov 30 17:49:19 PST 2025

Total time taken to generate the page: 2.48150 seconds