Project

General

Profile

Forums » Open discussion »

seg fault invoking splitseas on latest NOAA OLR data

Added by Kyle Clem over 3 years ago

Hi CDO users,

I've downloaded the latest NOAA interpolated OLR data from https://psl.noaa.gov/data/gridded/data.interp_OLR.html (attached below). I've worked with this dataset for many years using CDO. However, for some reason I get a segmentation fault invoking "splitseas" on this latest OLR update. I have tried with an older (1.6.3) and newer (1.9.5) CDO version. I can do basic "cdo info" and even basic calculations "cdo mulc,5", but not splitseas.

Anyone know why?

Thanks,
Kyle

olr.mon.mean.nc (14.6 MB) olr.mon.mean.nc

Replies (6)

RE: seg fault invoking splitseas on latest NOAA OLR data - Added by Brendan DeTracey over 3 years ago

What cdo version/platform? Works for me on Linux-Anaconda-CDO1.9.10

$ cdo -v splitseas olr.mon.mean.nc test.nc
 OpenMP:  num_procs=24  max_threads=1
cdo    splitseas: create file test.ncJJA.nc
cdo    splitseas: create file test.ncSON.nc
cdo    splitseas: create file test.ncDJF.nc
cdo    splitseas: create file test.ncMAM.nc
cdo    splitseas: Processed 5876208 values from 1 variable over 559 timesteps [0.32s 65MB].

RE: seg fault invoking splitseas on latest NOAA OLR data - Added by Karin Meier-Fleischer over 3 years ago

Hi Kyle,

just like Brendan, I have no problems using splitseas and your dataset on OS X 10.15 with CDO version 1.9.9.

-Karin

RE: seg fault invoking splitseas on latest NOAA OLR data - Added by Kyle Clem over 3 years ago

Hi Karin and Brendan,

Thank you both so much for your reply. Following your feedback, I have an update on my seg fault. It is only when I pipe "seasmean" and "splitseas" that I have a problem (as shown below), even though that is what I always do without problems

$ cdo splitseas -seasmean olr.mon.mean.nc test_
cdo splitseas: Started child process "seasmean olr.mon.mean.nc (pipe1.1)".
Segmentation fault

But if I do a "seasmean" or "splitseas" separately, it works just like it did for you. Strange. Are either of you able to pipe "seasmean" and "splitseas"?

Thanks,
Kyle

RE: seg fault invoking splitseas on latest NOAA OLR data - Added by Karin Meier-Fleischer over 3 years ago

Sorry, but I can't reproduce the problem.

cdo splitseas -seasmean olr.mon.mean.nc test_
cdo(1) seasmean: Process started
cdo(1) seasmean (Warning): Season 187 ( 2020-12-01) has only 1 input time step!
cdo(1) seasmean: Processed 5876208 values from 1 variable over 745 timesteps.
cdo    splitseas: Processed 1965744 values from 1 variable over 187 timesteps [0.35s 40MB].

ls -la test_*
-rw-r--r--  1 k204045  staff  2007458  2 Mär 09:59 test_DJF.nc
-rw-r--r--  1 k204045  staff  2007458  2 Mär 09:59 test_JJA.nc
-rw-r--r--  1 k204045  staff  1965410  2 Mär 09:59 test_MAM.nc
-rw-r--r--  1 k204045  staff  2007458  2 Mär 09:59 test_SON.nc

RE: seg fault invoking splitseas on latest NOAA OLR data - Added by Brendan DeTracey over 3 years ago

What is your platform? It works for me too:

$ cdo splitseas -seasmean olr.mon.mean.nc test_
cdo(1) seasmean: Process started
cdo(1) seasmean (Warning): Season 187 ( 2020-12-01) has only 1 input time step!
cdo(1) seasmean: Processed 5876208 values from 1 variable over 745 timesteps.
cdo    splitseas: Processed 1965744 values from 1 variable over 187 timesteps [0.44s 51MB].

Please try a more recent cdo version.

RE: seg fault invoking splitseas on latest NOAA OLR data - Added by Kyle Clem over 3 years ago

Hi Brendan and Karin,

It may indeed be the fact that I am using an older version of CDO. Regardless, I have an easy work around thanks to your help (simply do seasmean and splitseas separately). Strangely, if I convert to srv (cdo -f srv splitseas -seasmean olr.mon.mean.nc test_ ) this works perfectly fine too.

Thank you both so much for your quick replies and help!! :)

Best wishes,
Kyle

    (1-6/6)