A DSL-based implementation of the MPDATA weather kernel achieves 2.1x speedup over an OpenACC version on a K80 GPU, supporting the claim that a single source code can be performance portable.
It allows implementing halo exchanges and general boundary conditions (in multiple dimensions) that abstract away the architecture dependent implementation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Report on the performance portability demonstrated for the relevant Weather & Climate Dwarfs
A DSL-based implementation of the MPDATA weather kernel achieves 2.1x speedup over an OpenACC version on a K80 GPU, supporting the claim that a single source code can be performance portable.