comp.lang.idl-pvwave archive: archive

Home » Public Forums » archive » Why is MEAN so slow?

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Why is MEAN so slow? [message #74499]

Tue, 18 January 2011 16:52

Matt Francis
Messages: 94
Registered: May 2010

Member

I've been using the PROFILER to track down why some code was a bit
slow and found that it was spending most of its time in the MEAN
function (and then within that, in the MOMENT function called by
MEAN).

I had a look at the code for this and it looked very inefficient. It
took just moments to come up with this straight forward version:

function mean_quick,data,nan=nan,double=double

if keyword_set(nan) then begin
indx = where(finite(data),count)
if count EQ 0 then return, keyword_set(double) ? !values.D_nan : !
values.f_nan
return,total(data[indx],double=double)/count
endif else begin
return,total(data,double=double)/n_elements(data)
endelse

end

This is very simple minded, but as far as I can tell matches the
functionality of the library function MEAN and is at least 5 times
faster (for my usage in this code).

Why would the Library function be written so inefficiently?

Report message to a moderator

[Message index]

Why is MEAN so slow?

By: Matt Francis on Tue, 18 January 2011 16:52

Previous Topic:	How to speed up code which checks lots of values of an array
Next Topic:	FINDFILE problems IDL 8 LINUX

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Nov 29 10:23:45 PST 2025

Total time taken to generate the page: 0.72320 seconds