comp.lang.idl-pvwave archive: archive » Very tricky NetCDF "bug"

Home » Public Forums » archive » Very tricky NetCDF "bug"

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

Re: Very tricky NetCDF "bug" [message #82545 is a reply to message #82463]

Sat, 15 December 2012 09:29

Fabzi
Messages: 305
Registered: July 2010

Senior Member

Hi Chris,

On 12/15/2012 06:22 AM, Chris Torrence wrote:
> Presumably, within the NetCDF4 library, there are two different code
>paths, one for "no stride" (i.e. stride=1) and the other for stride=N.
>It's probably not checking to see if you pass in stride=1, but is just
> going straight to the "slower" stride code.

Yes, that's what I was thinking. It's not really "a bug", but it's just
that my process completely froze and it was really hard to find the
bottleneck...

> So, in short, I would recommend putting a check into your pro code,
> and only pass in the stride when you absolutely have to.
Yes, that's what I did.

Thanks,

Fab

Report message to a moderator

[Message index]

		Very tricky NetCDF "bug" By: Fabzi on Fri, 14 December 2012 10:08
		Re: Very tricky NetCDF "bug" By: Fabzi on Sat, 15 December 2012 09:29
		Re: Very tricky NetCDF "bug" By: chris_torrence@NOSPAM on Fri, 14 December 2012 21:22

Previous Topic:	Sparse matrix algorithms
Next Topic:	Coyote Library Update - Color Handling

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Oct 10 12:44:45 PDT 2025

Total time taken to generate the page: 0.15933 seconds