The Parallel Spectral
Transform Shallow Water Model (PSTSWM) is a message-passing benchmark
code and parallel algorithm testbed that solves the nonlinear shallow water
equations on a rotating sphere using the spectral transform method.
PSTSWM was developed by
P. H. Worley at Oak Ridge National Laboratory and
I. T. Foster at Argonne National Laboratory
from the serial code STSWM, written by J. J. Hack and R. Jakob at the National
Center for Atmospheric Research. PSTSWM was used to evaluate parallel
algorithms
for the spectral transform method as it is used in global atmospheric
circulation models, as part of the Department of Energy CHAMMP research
program. See
PSTSWM has some characteristics that have made it very useful
for performance studies:
Multiple parallel algorithms are embedded in the code and can be selected
at run-time, as can the problem size, number of processors, and data
decomposition.
Multiple message-passing transport layers are supported, including PVM,
MPI, and SHMEM.
Performance instrumentation is embedded in the code (using PICL), supporting both detailed
profiling and event tracing.
The code is an accurate representation of an important kernel
in spectral atmospheric models.
A partial bibliography of the performance studies utilizing PSTSWM can
be found here.
PSTSWM is also a "compact application" in the Parallel Kernels and Benchmarks
Suite (ParkBench).
The performance work continues, with the inclusion of new parallel
platforms and the examination of new performance questions.
To view the current status of the research, follow these links:
Worley's Performance Studies Page
Patrick H. Worley / (
worleyph@ornl.gov)