TY - GEN
T1 - Adapting Atmospheric Chemistry Components for Efficient GPU Accelerators
AU - Ruiz, Christian Guzman
AU - Dawson, Matthew
AU - Acosta, Mario C.
AU - Jorba, Oriol
AU - Galobardes, Eduardo Cesar
AU - García-Pando, Carlos Pérez
AU - Serradell, Kim
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2023
Y1 - 2023
N2 - Atmospheric models demand a lot of computational power, and solving the chemical processes is one of its most computationally intensive components. This work shows how to improve the computational performance of the Multiscale Online Nonhydrostatic AtmospheRe CHemistry (MONARCH), a chemical weather prediction system developed by the Barcelona Supercomputing Center. The model implements the new flexible external package chemistry across multiple phases (CAMP) for the solving of gas- and aerosol-phase chemical processes that allows multiple chemical processes to be solved simultaneously as a single system. We introduce a novel strategy to simultaneously solve multiple instances of a chemical mechanism, represented in the model as grid cells, obtaining a speedup up to 9$$\times $$ using thousands of cells. In addition, we present a GPU strategy for the most time-consuming function of CAMP. The GPU version achieves up to 1.2$$\times $$ speedup compared to CPU. Also, we optimize the memory access in the GPU to increase its speedup up to 1.7$$\times $$.
AB - Atmospheric models demand a lot of computational power, and solving the chemical processes is one of its most computationally intensive components. This work shows how to improve the computational performance of the Multiscale Online Nonhydrostatic AtmospheRe CHemistry (MONARCH), a chemical weather prediction system developed by the Barcelona Supercomputing Center. The model implements the new flexible external package chemistry across multiple phases (CAMP) for the solving of gas- and aerosol-phase chemical processes that allows multiple chemical processes to be solved simultaneously as a single system. We introduce a novel strategy to simultaneously solve multiple instances of a chemical mechanism, represented in the model as grid cells, obtaining a speedup up to 9$$\times $$ using thousands of cells. In addition, we present a GPU strategy for the most time-consuming function of CAMP. The GPU version achieves up to 1.2$$\times $$ speedup compared to CPU. Also, we optimize the memory access in the GPU to increase its speedup up to 1.7$$\times $$.
KW - Chemistry
KW - Parallelism and concurrency
KW - Performance
UR - https://www.scopus.com/pages/publications/85174738577
U2 - 10.1007/978-981-99-3091-3_11
DO - 10.1007/978-981-99-3091-3_11
M3 - Conference contribution
AN - SCOPUS:85174738577
SN - 9789819930906
T3 - Lecture Notes in Networks and Systems
SP - 129
EP - 138
BT - Proceedings of 8th International Congress on Information and Communication Technology - ICICT 2023
A2 - Yang, Xin-She
A2 - Sherratt, R. Simon
A2 - Dey, Nilanjan
A2 - Joshi, Amit
PB - Springer Science and Business Media Deutschland GmbH
T2 - 8th International Congress on Information and Communication Technology, ICICT 2023
Y2 - 20 February 2023 through 23 February 2023
ER -