Question about REMAP transform
Up to Bugs and debugs
Posted by Anonymous at January 7 2021
High,
I'm using the transform MAPPING of OASIS-MCT.
# Field 1 : Weighted sea surface temperature (o->a 1)
O_SSTSST SISUTESW 1 4 sstoc.nc
182 149 96 96 torc tlmd LAG=
P 2 P 0
LOCTRANS CHECKIN MAPPING CHECKOUT
# LOCTRANS: AVERAGE to average value over coupling period
AVERAGE
# CHECKIN: indicate computation of global, land and sea field integrals.
INT=1
# Mozaic: 1) mapping filename 2) connected unit 3) dataset rank 4) Maximum
# number of overlapped neighbors
rmp_torc_to_tlmd_MOSAIC.nc src
# CHECKOUT: indicate computation of global, land and sea field integrals.
INT=1
I provide a file with weights and addresses :
int src_address(num_links)
int dst_address(num_links)
double remap_matrix(num_links, num_wgts)
The result is a bit strange. The conservation of energy is not fully correct, and I need to check a few things :
Are src_address and dst_address in Fortran style (addresses from 1 to ...) or in C style (addresses from 0 to ...). My guess is Fortran style, but ?
What portions of the code uses theses weights and addresses ? I need to check that my weights computation is coherent with the way OASIS uses them.
Thanks, Olivier
Posted by Anonymous at January 8 2021
Laure, I finally find the error elsewhere ! Thank's for your help. Olivier
Posted by Anonymous at January 11 2021
Hi,
Yes the src_address and dst_address are in Fortran style (arrays index begins at 1).
Which version of OASIS3-MCT do you use ? OASIS3-MCT_2.0, OASIS3-MCT_3.0 ?
It will be easy to tell you where they are used if I know.
Thanks, Best regards, Laure
