| Re: Wavelet Coherency program [message #54249] |
Fri, 01 June 2007 03:13 |
tmgriff0206
Messages: 1
Registered: June 2007
|
Junior Member |
|
|
On 31 May, 22:20, "R.G. Stockwell" <n...@email.please> wrote:
> Have you calculated the number of points in the result? You might
> be running out of memory, and swapping to virtual memory.
> Perhaps you could try breaking the data into 14 time series of ~1024
> points each.
>
> Cheers,
> bob
Hi Bob,
Thanks for your reply, the number of points in the result is quite
large given it has the dimensions of number of points and scale (103
in this case) so that could be part of it. If I break the dataset
down into smaller chunks or reduce the time resolution so the runtime
is quicker then I loose information on the scales/periods that I'm
interested in so that's not really an option alas.
Thanks for your thoughts though,
T
|
|
|
|
| Re: Wavelet Coherency program [message #54252 is a reply to message #54249] |
Thu, 31 May 2007 14:20  |
news.qwest.net
Messages: 137
Registered: September 2005
|
Senior Member |
|
|
"tracy.moffat" <tracy.moffat@talk21.com> wrote in message
news:1180628991.572883.192550@o5g2000hsb.googlegroups.com...
> Hi,
>
> I'm using the wavelet coherency idl program provided by Torrence and
> Compo and it takes forever to run. I'm using quite small timeseries
> (14400 data points which can take up to 30minutes)
Have you calculated the number of points in the result? You might
be running out of memory, and swapping to virtual memory.
Perhaps you could try breaking the data into 14 time series of ~1024
points each.
Cheers,
bob
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
