-
Notifications
You must be signed in to change notification settings - Fork 133
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
Merge FMS dev/master into master #103
Conversation
…ing with 32-bit real
…n_fluxes_init with -O2 in ocean/ice_null configuratin. When compiled with -O2 using ifort 15.0.3 20150408 (Mac OSX), the model would crash "malloc: pointer being freed was not allocated" at the end of atmos_ocean_fluxes_init, at least in the configuration when ocean_null and ice_null are used (and of course there are no ocean fluxes). Some investigation shows that apparently part of the loop is evaluated ahead by the optimized code (even though the body is not supposed to be executed), and the strings constructed by concatenation are not properly allocated or deallocated by the compiler.
The `get_cpu_affinity` function in `mpp/affinity.c` had two return lines that essentially resulted in only one possible return value `first_cpu`. This commite reduces the return statement to simply `return first_cpu`. Resolves #62.
…-history-file' into dev/master
This update returns -1 if set_cpu_affinity has an error. The error message is removed to allow the calling routine to report the error.
Two void functions in affinity.c were given return values. Updated the function prototype to return an int.
The preprocessor macro MPP_DO_GLOBAL_FIELD_A2A_3D_ was not being cleared before being redefined for the various data types, and was raising warnings on some builds. This patch clears these macros before redefining them.
Merge in latest GDFL updates
…8 that only pass the optional data_default_r4 argument to setup_one_field if it is initally passed to register_restart_field
…ault, FLOAT_KIND)
Corrected a typo in setup_one_field 'compressed_axis' argument in register_restart_field_r2d8.
…gkindofdoctor/FMS into dev/master
The previous travis environment used an outdated openMPI the did not have a new MPI API element. The travis enviornment now uses xenial with an openMPI library with the missing API. Other travis.yml cleanup was also done. Also found that fm_util.o file was not included in the final libFMS library. This has been corrected. The base test did not link in the library, added a no-op libFMS call to constants_init(), and modified the test Makefile.am to link in the library.
…_master_to_master_merge Resolved conflicts in affinity.c between changes in pull request #63 and version of affinity.c with apple #ifdefs
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.
I see nothing stopping adding these to master - assuming we have a release tag on master at the point these will be merged.
@bensonr No new changes--I just need approval to re-submit the pull request because the gitlab CI jobs timed out during the first attempt. |
@underwoo FYI, the gitlab-CI jobs are hanging before timing out with the following message:
|
This branch contains the most recent dev/master commits and affinity.c updates referenced in pull request #63 to merge into master so that users and developers can submit pull requests and issues directly to the master branch. Once this branch is merged, I will close #63 with a note that the affinity.c modifications have been applied in this pull request.