[MITgcm-support] coupler time averages writing frequency

[David Ferreira]

It's not usual, but it happens. Maybe try to create smaller nc files 
(with less fields).
david


On 1/13/14 12:24 PM, Alexandre Pohl wrote:
> Hi David,
>
> Thank you for your answer, it is exactly what I was looking for.
>
> Unfortunately, my problem persists even without the coupler time averages: the model becomes very slow as soon as netcdf outputs are created. Is it a usual problem? Does anybody know a way to make it run faster?
>
> Thank you in advance,
> Alexandre
>
>
>
>
> Le 11 janv. 2014 à 18:00, mitgcm-support-request at mitgcm.org a écrit :
>
>> Send MITgcm-support mailing list submissions to
>> 	mitgcm-support at mitgcm.org
>>
>> To subscribe or unsubscribe via the World Wide Web, visit
>> 	http://mitgcm.org/mailman/listinfo/mitgcm-support
>> or, via email, send a message with subject or body 'help' to
>> 	mitgcm-support-request at mitgcm.org
>>
>> You can reach the person managing the list at
>> 	mitgcm-support-owner at mitgcm.org
>>
>> When replying, please edit your Subject line so it is more specific
>> than "Re: Contents of MITgcm-support digest..."
>>
>>
>> Today's Topics:
>>
>>    1. coupler time averages writing frequency (Alexandre Pohl)
>>    2. Re: coupler time averages writing frequency (David Ferreira)
>>    3. flt pkg (Ewa Katarzyna Karczewska)
>>    4. Re: flt pkg (Matthew Mazloff)
>>
>>
>> ----------------------------------------------------------------------
>>
>> Message: 1
>> Date: Fri, 10 Jan 2014 18:35:24 +0100
>> From: Alexandre Pohl <alexandre.pohl at lsce.ipsl.fr>
>> To: mitgcm-support at mitgcm.org
>> Subject: [MITgcm-support] coupler time averages writing frequency
>> Message-ID: <806C2254-4C73-49AA-A29A-DBA022DDA61F at lsce.ipsl.fr>
>> Content-Type: text/plain; charset=windows-1252
>>
>> Hi all,
>>
>> I am running the coupled MITgcm, and I have problems with the outputs. I first tried to use the meta/data output format, but the huge number of files that it produces exceeds my limitations in only 10 model-years (> 2Mo inodes).
>>
>> To solve my problem, I then decided to switch to MNC (netcdf) outputs. This time, the number of produced files is correct for me, but the model is so slow!
>>
>> I think that in each case, the main problem is the high writing frequency for the coupler time-averaged outputs. It creates a large number of meta/data files, and requires a lot of writing time when running the model with netcdf outputs.
>>
>> I am thus wondering if there is a way to reduce the coupler time averages writing frequency. I found no similar flag, contrary to the ocean and atmosphere that have this option in their input/data file? Or maybe somebody has another solution?
>>
>> Thank you in advance,
>> Alexandre
>>
>>
>> ------------------------------
>>
>> Message: 2
>> Date: Fri, 10 Jan 2014 20:10:06 +0000
>> From: David Ferreira <dfer at mit.edu>
>> To: mitgcm-support at mitgcm.org
>> Subject: Re: [MITgcm-support] coupler time averages writing frequency
>> Message-ID: <52D0539E.3060802 at mit.edu>
>> Content-Type: text/plain; charset=windows-1252; format=flowed
>>
>> Hi Alexandre,
>> Are out talking about the cpl_tave.*.t001.nc files ?
>> If so, look at cpl_taveFreq in  rank_1/data.cpl
>> cheers,
>> david
>>
>>
>> On 1/10/14 5:35 PM, Alexandre Pohl wrote:
>>> Hi all,
>>>
>>> I am running the coupled MITgcm, and I have problems with the outputs. I first tried to use the meta/data output format, but the huge number of files that it produces exceeds my limitations in only 10 model-years (> 2Mo inodes).
>>>
>>> To solve my problem, I then decided to switch to MNC (netcdf) outputs. This time, the number of produced files is correct for me, but the model is so slow!
>>>
>>> I think that in each case, the main problem is the high writing frequency for the coupler time-averaged outputs. It creates a large number of meta/data files, and requires a lot of writing time when running the model with netcdf outputs.
>>>
>>> I am thus wondering if there is a way to reduce the coupler time averages writing frequency. I found no similar flag, contrary to the ocean and atmosphere that have this option in their input/data file? Or maybe somebody has another solution?
>>>
>>> Thank you in advance,
>>> Alexandre
>>> _______________________________________________
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>>
>>
>>
>> ------------------------------
>>
>> Message: 3
>> Date: Sat, 11 Jan 2014 15:59:42 +0000
>> From: Ewa Katarzyna Karczewska <ekk29 at cam.ac.uk>
>> To: mitgcm-support at mitgcm.org
>> Subject: [MITgcm-support] flt pkg
>> Message-ID: <52D16A6E.2 at cam.ac.uk>
>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>
>> Dear all,
>>
>> I'm using the flt package in an offline run.
>>
>> I've noticed that the float_trajectories.xxx.xxx.data output files are put into the hidden
>> subdirectories (.0xxx/ )in my /run folder. Why is that? Which setting
>> should I change so that the subdirectories the output files are written
>> to are not hidden?
>>
>> Also, I'm getting .meta files only for some of the output files
>> (float_trajectories.xxx.xxx.data), can I change it?
>>
>> Please let me know if you need more detailed information on how I set my
>> parameters etc.
>>
>> Best,
>> Ewa
>>
>>
>>
>>
>>
>>
>> ------------------------------
>>
>> Message: 4
>> Date: Sat, 11 Jan 2014 08:28:05 -0800
>> From: Matthew Mazloff <mmazloff at ucsd.edu>
>> To: <mitgcm-support at mitgcm.org>
>> Subject: Re: [MITgcm-support] flt pkg
>> Message-ID: <F554A022-9E83-4E63-B15B-0C055CB2CDAE at ucsd.edu>
>> Content-Type: text/plain; charset=windows-1252
>>
>> Hi Ewa
>>
>> Regarding .000 directories, is this only for float diagnostic output or for all output?
>> and make sure   mdsioLocalDir   is not set to anything in your data file?if you have it set to
>>   mdsioLocalDir = '.',
>> that would explain what you are getting.
>>
>> Regarding .metas -- is it only some timesteps, or is it only some types of output?
>>
>> Finally -- have you made any changes to the default code -- the relevant routines would likely be
>> ini_model_io.F
>> mdsio_write_meta.F
>>
>> -Matt
>>
>>
>>
>>
>>
>> On Jan 11, 2014, at 7:59 AM, Ewa Katarzyna Karczewska <ekk29 at cam.ac.uk> wrote:
>>
>>> Dear all,
>>>
>>> I'm using the flt package in an offline run.
>>>
>>> I've noticed that the float_trajectories.xxx.xxx.data output files are put into the hidden
>>> subdirectories (.0xxx/ )in my /run folder. Why is that? Which setting
>>> should I change so that the subdirectories the output files are written
>>> to are not hidden?
>>>
>>> Also, I'm getting .meta files only for some of the output files
>>> (float_trajectories.xxx.xxx.data), can I change it?
>>>
>>> Please let me know if you need more detailed information on how I set my
>>> parameters etc.
>>>
>>> Best,
>>> Ewa
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>>
>>
>>
>> ------------------------------
>>
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>>
>>
>> End of MITgcm-support Digest, Vol 127, Issue 12
>> ***********************************************
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support