modern_diag_manager: fms_diag_do_reduction #1321
Conversation
…arguments are correct
| LOGICAL, DIMENSION(:,:,:,:), pointer, INTENT(in) :: mask !< The location of the mask | ||
| CLASS(*), DIMENSION(:,:,:,:), pointer, INTENT(in) :: rmask !< The masking values |
There was a problem hiding this comment.
do these have to be pointer? Is it just because their input arguments are pointers? Is that a requirement in Fortran? Since Fortran is pass by reference, I don't think these need to be pointers.
There was a problem hiding this comment.
It is required because I am doing if(associated(mask))
|
|
||
| real(kind=r8_kind) :: out_weight | ||
|
|
||
| out_weight = 1.0_r8_kind |
There was a problem hiding this comment.
Did you run this by @mlee03 ?
There was a problem hiding this comment.
1.0_r8_kind is preferred over real(1.0 kind=r8_kind)
| type is (real(kind=r8_kind)) | ||
| out_weight = real(weight, kind = r8_kind) | ||
| type is (real(kind=r4_kind)) | ||
| out_Weight = real(weight, kind = r4_kind) |
There was a problem hiding this comment.
You declared out_Weight as r8, but here you are making it r4? Maybe you need out_Weight to be class(*) and allocate it to whatever weight is? Then out_weight = 1.0_r8_kind would be in the class default branch of the select type. Otherwise you should have kind=r8_kind on this line I think.
There was a problem hiding this comment.
out_weight is going to be saved as a r8_kind and converted to r4_kind if needed when doing the mad.
That should be kind=r8_kind no kind=r4_kind
|
@thomas-robinson I think everything was addressed in 91a249a |
| field_modern(1:ie,1:je,1:ke,1:1) => field | ||
| field_remap(1:ie,1:je,1:ke,1:1) => field | ||
| if (present(mask)) mask_remap(1:ie,1:je,1:ke,1:1) => mask | ||
| if (present(rmask)) rmask_remap(1:ie,1:je,1:ke,1:1) => rmask |
There was a problem hiding this comment.
Why isn't rmask declared with contiguous?
There was a problem hiding this comment.
I am not sure why the compiler didn't complain, but i declared it ascontiguous
|
@uramirez8707 There is a merge conflict |
|
@thomas-robinson the "merge conflict" was resolved |
Description
This contains part of #1295
Fixes # (issue)
How Has This Been Tested?
CI
Checklist:
make distcheckpasses