How Can I Define "Enable PerfCheck" in Namelist.rc about regesm

[smft]

I compiled all the codes with gnu compiler. It seems that the code is not compatible with intel compile version 2018 update 2 perfectly.

Is regesm compiler sensitive?

[uturuncoglu]

Hi, That is great! Actually, the version of RegESM that is used in the workshop is little bit old. It might be compatible with RegESM 1.0. After that we introduced co-processing component (in 1.1) and EnablePerfCheck is an option to measure wall clock time in predefined regions in the code when co-processing component enabled. So, if you plan to use master, please include that option to the namelist.rc (I think that you are using namelist.rc coming from workshop repository). There might be also other differences but I am not sure. Just try and let me know when you need help. In addition, after RegESM 1.1 the modeling system uses Markdown based documentation rather than static PDF file. It is more convenient way to keep documentation up to date. You can find the link for the specific section of documentation in the README.md in the RegESM repository page. For example, installation section is in installation and running. For the question related with compiler, there is no any dependency. I could successfully compiled model with Intel and gnu compilers but again it might change based on your working environment and used version of the required libraries. So, let me know when you have a problem also with installation. I think you are very close to have a working system.

[smft]

Thanks a lot ! RegESM is running in a concurrent mode, while sequential is not working. I think this maybe results from my environment settings. One more thing, I find that regesm's default Coupling Type is ocean -------> atmosphere (explicit) and ATM------->OCN is implicit. But I do not quite understand the other two mode 'Semi-implicit' and 'leap frog' listed in your namelist.rc. Any references?
Many Thanks!

[uturuncoglu]

you might find more information about sequential vs. concurrent and their setup in the Markdown documentation. In semi-implicit the ocean model is one step ahead. So, there is a one coupling time step lag between them. Following figure could demonstrate it,

Fig03.pdf

This is a convention in ESMF side, so to get more information you could check here.

[smft]

Roger that, many thanks and have a nice day!

…

在 2018年4月25日，下午6:27，Ufuk Turunçoğlu ***@***.***> 写道： you might find more information about sequential vs. concurrent and their setup in the Markdown documentation. In semi-implicit the ocean model is one step ahead. So, there is a one coupling time step lag between them. Following figure could demonstrate it, Fig03.pdf This is a convention in ESMF side, so to get more information you could check here. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.