Re: memory size calculation

[Ansley Manke <Ansley.B.Manke@xxxxxxxx>]

Hello Hein,
The erorr message may be reporting only part of the actual
memory that was reqired for the command to run.  There
are a couple of discussions of memory management and
requirements in the Users Guide.  Please look under these
two index entries

   "memory, insufficient memory"
and
  "memory, large calculations"

In the discussion of large calculations, you will see that when
you have made a number of variable definitions, the size of
the data that must be loaded to evaluate them may be far
larger than the size of the final result.  You may be able to
see what is happening by setting SET MODE DIAGNOSTIC
before your SAVE command.  There is an FAQ on how to
interpret the output of the diagnostic mode, at

http://ferret.pmel.noaa.gov/Ferret/FAQ/other/mode_diagnostic.html

Please let us know if you still have questions about the use
of memory in Ferret.

Ansley Manke

Hein Zelle wrote:

> Why does the following happen?
>
> set memory/size=50
>  Cached data cleared from memory
>
> save /clobber /file=wdtdz.nc wprimedtdz,wdtdzprime,wprimedtdzprime
>  **ERROR: insufficient memory: 10931760 words were requested.
>
> As far as I can tell, I've just requested ferret to allocate 50
> megawords of memory, and yet it complains that it needs 11 megawords
> and doesn't have enough. Hmmz? When I increase memory to 100, it
> works.
>
> Hein Zelle
>
> >-----------------------------------------------------------------<
>     Hein Zelle
>     Dept. of Oceographic Research
>     KNMI, The Netherlands
>     work:        zelle@knmi.nl     http://www.knmi.nl/~zelle
>     private:     hein@icce.rug.nl  http://www.icce.rug.nl/~hein
>     Phone:       +31 (0)30 2206704
> >-----------------------------------------------------------------<
> --------------------------------------------------------------
> Zie ook/see also: http://www.knmi.nl/maildisclaimer.html

--
Ansley Manke  Pacific Marine Environmental Laboratory  Seattle WA