|
|
PCCM2.1 is a parallel implementation of version 2.1 of the CCM developed in the early to mid 1990s for the Department of Energy CHAMMP research program by a collaboration of researchers from Argonne and Oak Ridge National Laboratories and NCAR. It uses a two-dimensional domain decomposition approach that allows up to 1024 processes to be used for T42L18. PCCM2.1 was originally targeted for the Intel Paragon with 1024 processors and the IBM SP2 with 128 processors. But the code was written to be easily ported to other multiprocessors that support message-passing paradigms, or to run on machines distributed across a network with PVM.
PCCM2.1 was not adopted as the production CCM by NCAR because CCM had reached version 3.0 by the time PCCM2.1 was validated. There were also perceived difficulties in maintaining efficiency on parallel vector processors when incorporating the code modifications for the two-dimensional decomposition. However, the computational requirements of the new generation of climate models, as constrained by the parallel platforms currently available to U.S. climatologists, may require the use of large numbers of processors in order to achieve the goals of the U.S. Global Change program. To assess the potential for performance gains from exploiting additional parallelism in the atmospheric model, PCCM was updated (by John Drake and Pat Worley of ORNL and Rodney James of NCAR) to use the same physics routines and numerical methods as version 3.2 of CCM. As the current version of CCM is also a parallel code, PCCM is no longer a useful name for the two-dimensional parallelization. We refer to the new version of PCCM as CCM/MP-2D.
Patrick H. Worley / (
worleyph@ornl.gov)