ABSTRACT: Numerical Weather Prediction (NWP) models are, at first glance, not ideal
candidates for GPUs: they tend to be memory bandwidth limited and have
calculation intensities of only few floating point operations per
datum in memory. Upon closer examination, though, many of the
employed algorithms operate on vertical 'columns' in the underlying
three dimension fields which have no neighbor dependencies. This
attribute allows the calculation to be formulated in thread blocks.
Furthermore, these models benefit from the much higher memory
bandwidth of GPUs.
There is a nascent community of atmospheric modelers which is looking
toward GPUs as a stepping stone to higher resolution modeling which
will enable new scientific development. Already several full-up
models have been successfully ported to GPUs. But up to
today, there has been few community discussions about the issues in
programming GPUs for NWP.
Our objectives for this Birds-of-a-Feather are:
(1) to have at least two teams present there
experiences in porting NWP models to GPUs, (2) to foster the evolving
community of atmospheric modelers who calculate on, or would like to
calculate on, GPUs, (3) to exchange information about programming
techniques, code performance and obstacles to parallelization, and
(4) to share ideas on how the task of programming NWP models on GPUs
can be facilitated by hardware and software vendors.
Session Leader Details:
William Sawyer (Primary Session Leader) - Swiss National Supercomputing Centre