Re: [netcdfgroup] read performance slow compared to netCDF on other systems

  • To: "Thompson, Matt (GSFC-610.1)[SCIENCE SYSTEMS AND APPLICATIONS INC]" <matthew.thompson@xxxxxxxx>
  • Subject: Re: [netcdfgroup] read performance slow compared to netCDF on other systems
  • From: Liam Forbes <loforbes@xxxxxxxxxx>
  • Date: Fri, 11 Nov 2016 07:56:59 -0900
On Fri, Nov 11, 2016 at 5:22 AM, Thompson, Matt (GSFC-610.1)[SCIENCE
SYSTEMS AND APPLICATIONS INC] <matthew.thompson@xxxxxxxx> wrote:

> A question: are you using the same version of NCO in both cases? A while
> back (year or two) I encountered a performance issue that was due to NCO
> and not netCDF (I tried many a combination). Charlie fixed whatever was
> causing the slowdown a while back, so if you just built NCO it's probably
> not that, but I wanted to make sure.
>

Matt,

I've built and am using NCO 4.6.1 on our new cluster. The researchers who
reported the problem aren't using NCO commands, they are using NetCDF from
within their applications. I hadn't considered that NCO might be skewing my
test case though. Is there another way to do whatever ncks is doing when it
reads in a file and writes out a new one with no options specified on the
command line? I can see that it's not just an nccopy because the resulting
file is different. I guess this is where my ignorance of what these
packages do starts to show.

Our comparison case, the Cray, has NCO 4.1.0.

-- 
Regards,
-liam

-There are uncountably more irrational fears than rational ones. -P. Dolan
Liam Forbes  loforbes@xxxxxxxxxx  ph: 907-450-8618 fax: 907-450-8601
UAF Research Computing Systems Senior HPC Engineer  LPIC1, CISSP
  • 2016 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdfgroup archives: