Re: CRAY version

[hankin@xxxxxxxxxxxxxxxxxxx (Steve Hankin)]

Hi Mark,

The solution that I would strongly advocate for interactions between Ferret and
your Cray is to NFS-export the Cray volumes on which your data is stored and
NFS-mount them on your workstation. No file transfers then need be done. Since
Ferret uses random access when reading the data (reading chunks of modest size)
the NFS connection is typically very satisfactory performance.

(Hopefully your Cray management folks are cooperative about NFS exports. You
could point out to them that a read-only export would be fine.)

As to running Ferret directly on the Cray, well, Ferret would be able to take
little advantage of array processing capabilities so it would be "expensive" to
run. And, as you pointed out, porting Ferret is a signiificant chunk of work.
So unless we hear a great outcry, a port of Ferret to Cray is not very likely.


	- steve

                |  NOAA/PMEL               |  ph. (206) 526-6080
Steve Hankin    |  7600 Sand Point Way NE  |  FAX (206) 526-6744
                |  Seattle, WA 98115-0070  |  hankin@pmel.noaa.gov

=====================================================================

On Apr 18, 11:18am, Mark Collier wrote:
> Subject: CRAY version
> G'day Ferreters,
> is there any likelyhood that a CRAY version of ferret will be
> released?
>
> It seems that most earth scientists have access to a supercomputer
> (I guess usually a CRAY of some form) and ferret would be an ideal
> platform to run it on. Secondly, supercomputers usually mean the
> production of large datasets and the availability of ferret could
> reduce the need to transfer data to small workstations and provide
> an efficient way in monitoring a production run. I personally find
> ferret to be a good means to instantly spot suspicious data in a large
> data set (for whatever reason).
>
> I know porting software is a difficult process but just wondering
> if something was in the pipeworks? Regards, Mark.
> *=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*
> * CRC for SHM and CSIRO DAR: Research Assistant/Programmer            *
> *    e-mail: mark.collier@dar.csiro.au                                *
> *     voice: (03) 9239 4420 +61-3-9239 4420                           *
> *     snail: PMB 1 Aspendale, Victoria, Australia, 3195.              *
> *=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*
>
>-- End of excerpt from Mark Collier




-- 

		|  NOAA/PMEL               |  ph. (206) 526-6080  
Steve Hankin	|  7600 Sand Point Way NE  |  FAX (206) 526-6744
		|  Seattle, WA 98115-0070  |  hankin@pmel.noaa.gov