-
Notifications
You must be signed in to change notification settings - Fork 146
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Gmtb gfsphysics gnu support #36
Gmtb gfsphysics gnu support #36
Conversation
…b.F: gfortran crashes when 'real (kind=kind_io8) data8(idim*jdim)' is used, i.e. turn into allocatable arrays
…) Modifications of makefile to handle different name and order of parts of name for shared physics library on MacOSX; add missing CPPDEFS argument to compile commands
…ix' arrays to handle pointers to components of derived-type arrays (see also following commits to GFS diagnostics and fms I/O code)
…ys in GFS_diagnostics.F90 to deal with issue of pointers to components of derived-type arrays
… into gmtb-gfsphysics-gnu-support
These changes look reasonable to me. I will run a quick ifort/cheyenne test next. Julie - are you going to be able to test this version with gfortran? |
Good luck with Cheyenne - it's slow as right now. Started compiling at 9am this morning, thus far it completed 5 fortran source files ... and Theia isn't available either today! Time to get these Mac OS X changes in the PR in ;-) |
Hi Laurie,
The two platforms I have available are theia, which is down today and
features GNU4.something. I also have a Docker container from the JEDI
project, which I haven't looked at in a while. I don't have an account on
cheyenne. But it is a great day to go for a run, 65 F in January!
Julie
…On Tue, Jan 9, 2018 at 2:09 PM, Laurie Carson ***@***.***> wrote:
These changes look reasonable to me. I will run a quick ifort/cheyenne
test next.
Julie - are you going to be able to test this version with gfortran?
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#36 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ADdm00rX0Fi-1EDcARaZCjUNFEhfZhXAks5tI9WigaJpZM4RW-Fn>
.
|
Good points, all! I've been going back-and-forth with CISL all day about
cheyenne "it's normal to have slow file system access..." yeah, right. It
took me over an hour to run the configure script for WRF (which compiles 2
tiny test programs). !!! At least this afternoon, they (CISL)
acknowledged that it's not "normal".
Enjoy your run, Julie!
On Tue, Jan 9, 2018 at 2:23 PM, JulieSchramm <notifications@github.com>
wrote:
… Hi Laurie,
The two platforms I have available are theia, which is down today and
features GNU4.something. I also have a Docker container from the JEDI
project, which I haven't looked at in a while. I don't have an account on
cheyenne. But it is a great day to go for a run, 65 F in January!
Julie
On Tue, Jan 9, 2018 at 2:09 PM, Laurie Carson ***@***.***>
wrote:
> These changes look reasonable to me. I will run a quick ifort/cheyenne
> test next.
>
> Julie - are you going to be able to test this version with gfortran?
>
> —
> You are receiving this because your review was requested.
> Reply to this email directly, view it on GitHub
> <#36 (comment)
>,
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/ADdm00rX0Fi-
1EDcARaZCjUNFEhfZhXAks5tI9WigaJpZM4RW-Fn>
> .
>
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#36 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AHTrItIeii0ZFV-g78omxILe3omJxHsqks5tI9i_gaJpZM4RW-Fn>
.
|
@JulieSchramm @llpcarson On Theia gnu-6.2.0 is used. The compiler is provided as a module, the remaining software stack I compiled by myself. The module file will add the correct directories to the PATH and LD_LIBRARY_PATH, and will set other environment variables (e.g. NETCDF). So, you can use Theia once it is back. If you use docker, you will have to provide the software stack (compiler, mpi library, netcdf library, esmf library) by yourself. |
Gmtb ccpp gnu support - requires NCAR/ccpp-physics#36 followed by NCAR/gmtb-fv3#25 for full support of gcc/clang compilers.
@llpcarson If you are happy with it, can you please approve this one as well? Needs to go in before the gmtb-fv3 gnu support PR (https://github.com/NCAR/gmtb-fv3/pull/25). Thanks! Dom |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Approved
Done - sorry, I thought I had approved the submodule ones already!
…On Fri, Jan 12, 2018 at 8:58 AM, Dom Heinzeller ***@***.***> wrote:
@llpcarson <https://github.com/llpcarson> If you are happy with it, can
you please approve this one as well? Needs to go in before the gmtb-fv3 gnu
support PR (NCAR/gmtb-fv3#25 <https://github.com/NCAR/gmtb-fv3/pull/25>).
Thanks! Dom
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#36 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AHTrIlpJLbqHSXXwrWKQi376AYj3pzn5ks5tJ4D6gaJpZM4RW-Fn>
.
|
MYNN PBL and SFC updates for ocean coupling, ...
This PR enables support for (reasonably recent) GNU compilers gcc+gfortran on Theia and Cheyenne, as well as for clang+gfortran on MacOSX. Besides several bugs in the code that are addressed in this PR, a temporary workaround is implemented for issues with pointers to components of pointer arrays of derived data types. These work with Intel and PGI compilers, but not with GNU compilers. Limited research on the web shows that this i a long-standing bug in the GNU compilers.