[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 20050111: Performance problem with NetCDF-3.5

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: 20050111: Performance problem with NetCDF-3.5
  • Date: Wed, 19 Jan 2005 09:13:43 -0700
>To: <address@hidden>
>From: "Christelle Riou" <address@hidden>
>Subject: Performance problem with NetCDF-3.5
>Organization: Capgemini Group
>Keywords: 200501110902.j0B92tv2004020 netCDF performance

Hi Christelle,

> I encounter performance problem in using the NetCDF-3.5 library.
> 
> When i add a large variable in a file, it takes not very long, but
> the next "get" instruction is very long even if I force the
> data_mode before the "get" instruction. The "get" instruction seems
> to move variables and that takes very long time.
> 
> What can I do to avoid this ?

The "get" functions in netCDF don't move any bytes in the file.

They might take a long time if you are getting values that are not
close to each other on disk.  For example, if you have written a large
2D variable by rows and are accessing values by column, it is possible
that each value could require a separate disk access, which would be
slow.

Do you have a small example that would allow us to duplicate the
problem here, so we could suggest a solution?

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden          http://www.unidata.ucar.edu/staff/russ