search

Urban-Meteorology-Reading / WRF-SUEWS

WRF-SUEWS coupling project
https://wrf-suews.readthedocs.org
MIT License
5 stars 2 forks source link

`ERROR reading sector coeff of namelist.suews` #43

Closed sunt05 closed 5 years ago

sunt05 commented 5 years ago

@hamidrezaomidvar Have you modified the structure of namelist.suews in recent days? I meet the same problem as before when I run wrf.exe. I use the version you cloned for me.

ERROR reading sector coeff of namelist.suews

Originally posted by @zhenkunl in https://github.com/Urban-Meteorology-Reading/WRF-SUEWS/issues/37#issuecomment-496636138

sunt05 commented 5 years ago

https://github.com/Urban-Meteorology-Reading/WRF-SUEWS/issues/37#issuecomment-496667715

sunt05 commented 5 years ago

@hamidrezaomidvar can you remind me of the cause and fix for this error when you encountered it last time?

zhenkunl commented 5 years ago

That is quite strange. I think it isn't the the matter of file permission. I changed the permission of namelist.suews as well, though it didn't work.

zhenkunl commented 5 years ago

Also can you set the variables in method sector to integer other than real? wrf.exe complains about that.

hamidrezaomidvar commented 5 years ago

Last time that I was encountering this problem were mostly because of MPI problem. I was running multiple runs at the same time and for the new ones I was getting this error. Maybe just for test, decrease the number of processors to see if this is the problem. I have run multiple cases with the latest version, and it was fine!

On May 28, 2019, at 9:11 PM, Ting Sun notifications@github.com wrote:

@hamidrezaomidvar can you remind me of the cause and fix for this error when you encountered it last time?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

hamidrezaomidvar commented 5 years ago

The param extractor program already takes care of it! Are you using the latest version? Double check this since it should be integer automatically

On May 28, 2019, at 9:24 PM, Li Zhenkun notifications@github.com wrote:

Also can you set the variables in method sector to integer other than real? wrf.exe complains about that.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

zhenkunl commented 5 years ago

It seems that the model start to run normally after I reduce the processors from 128 to 80. I remember you have told me about the type conversion, but I don't know why they appear in real type. I use the version you downloaded for me last week. From the git log, I can see

commit 1e0f16a80efbe7ab0967d3bcf0b5daf1bb06507d Author: Hamidreza Omidvar omidvar.hamidreza@gmail.com Date: Thu May 23 09:52:52 2019 +0100

changing the namelist.suews.new to namelist.suews
hamidrezaomidvar commented 5 years ago

Great! Let’s have a brainstorming session tomorrow to figure out if this MPI problem is From our side or Jasmin!

On May 28, 2019, at 9:40 PM, Li Zhenkun notifications@github.com wrote:

It seems that the model start to run normally after I reduce the processors from 128 to 80. I remember you have told me about the type conversion, but I don't know why they appear in real type. I use the version you downloaded for me last week. From the git log, I can see

commit 1e0f16a Author: Hamidreza Omidvar omidvar.hamidreza@gmail.com Date: Thu May 23 09:52:52 2019 +0100

changing the namelist.suews.new to namelist.suews — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

sunt05 commented 5 years ago

It seems that the model start to run normally after I reduce the processors from 128 to 80.

Just for test: may try to increase the number of processors and see when a similar error would be triggered again. so we might have a better clue of the issue.

zhenkunl commented 5 years ago

I didn't have dinner until now. Let's check it tomorrow as it is still running now

hamidrezaomidvar commented 5 years ago

Potential problem based on my discussion with @zhenkunl: the file unit number is being used by other processor in other part of WRF-SUEWS. So we can change the unit number to an unused one. @zhenkunl Can you please tell me if the test was successful based on the solution we discuss?

zhenkunl commented 5 years ago

The job is still running so I haven't conducted the test. Once it finishes, I will recompile the model to see if it can be solved. In addition, I glanced through the WRF code this morning and found the same file unit is used in different modules other than 10 (10 is used in our subroutine). So we should make it clear tomorrow.

zhenkunl commented 5 years ago

I have performed a test with a file unit 123 which is not used by other file, but it didn't work. The real reason of this problem is the parameter's type mismatch, as mentioned before, some integer parameters were assigned real values. @hamidrezaomidvar Is that a matter of library version of python?

suegrimmond commented 5 years ago

This is very worrying – and may mean that all results up to now are not correct. Great it we have found the problem (and actually good that there is a problem)

Best wishes Sue

Prof Sue Grimmond Dept. of Meteorology, University of Reading, Reading, RG6 6BB T: 44 118 378 6248 – messages get emailed to me O:Met Building (#58 on map) rm:1U14 E: c.s.grimmond@reading.ac.ukmailto:c.s.grimmond@reading.ac.uk W: http://micromet.reading.ac.uk/

From: Zhenkun Li notifications@github.com Sent: 30 May 2019 14:50 To: Urban-Meteorology-Reading/WRF-SUEWS WRF-SUEWS@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: Re: [Urban-Meteorology-Reading/WRF-SUEWS] ERROR reading sector coeff of namelist.suews (#43)

I have performed a test with a file unit 123 which is not used by other file, but it didn't work. The real reason of this problem is the parameter's type mismatch, as mentioned before, some integer parameters were assigned real values. @hamidrezaomidvarhttps://github.com/hamidrezaomidvar Is that a matter of library version of python?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/Urban-Meteorology-Reading/WRF-SUEWS/issues/43?email_source=notifications&email_token=AE2KZXQC74L4A3KH5ZXDFTLPX7LR3A5CNFSM4HQGVZW2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWSL7OQ#issuecomment-497336250, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AE2KZXV7BOGAJUC2I2IJX5TPX7LR3ANCNFSM4HQGVZWQ.

hamidrezaomidvar commented 5 years ago

@suegrimmond As I discussed with @zhenkunl , this problem was addressed before in the python script that I wrote where the namelist is being produced. For some reason that we still do not know, for @zhenkunl, some of the methods and coeff that need to be integer are written in real although it is addressed in the python script. Nevertheless, if they are written in real, the code would not run as there would be an error in reading them.

hamidrezaomidvar commented 5 years ago

@zhenkunl : can you please let me know the version of your python, pandas and numpy?

hamidrezaomidvar commented 5 years ago

@zhenkunl Can you try the modified script for SUEWS parameter extractor to see if this problem is fixed for you? You can find it here

zhenkunl commented 5 years ago

Sorry for late reply. I have tested the script and it does work well. BTW, the versions are (just on your request): pandas 0.23.4 numpy 1.15.4 Python 3.7.3