Japan Geoscience Union Meeting 2022

Presentation information

[E] Poster

P (Space and Planetary Sciences ) » P-PS Planetary Sciences

[P-PS03] Small Solar System Bodies: Latest results and new perspectives on the Solar System evolution

Thu. Jun 2, 2022 11:00 AM - 1:00 PM Online Poster Zoom Room (4) (Ch.04)

convener:Tatsuaki Okada(Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency), convener:Daisuke Kuroda(Kyoto University), Arika Higuchi(University of Occupational and Environmental Health, Japan), Chairperson:Daisuke Kuroda(Kyoto University), Arika Higuchi(University of Occupational and Environmental Health, Japan), Tatsuaki Okada(Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency)

11:00 AM - 1:00 PM

[PPS03-P07] Introducing dynamic parallel-processing in HEAT

*Ryuji Konno1, Saito Hiroto1, Takehiko Arai2, Hirohide Demura1 (1.The University of Aizu, 2.Maebashi Institute of Technology)

Keywords:software, Hayabusa2, Hera

HEAT (Hayabusa2/Hera Exploration Assistant for TIR/TIRI) is a calibration tool for thermal imagers; TIR on Hayabusa2 and TIRI on Hera. These imagers observe infrared radiation as digital number. These main functions are to know the nature of the asteroid based on thermo-physical properties. HEAT has been developed by a team of University of Aizu through missions of Hayabusa2 and Hera. HEAT has four functions, Calibration, DB, Visualization, and Conversion. Target of this research is introduction of parallel processing to the calibration process as improvement of its performance. HEAT calibrates 328 * 248 (=81,344) pixels one by one. These repeated tasks need parallel processing for reducing processing time. Suko+(2020) implemented a way of parallel processing designed for only 6 core system. This hard cording brings insufficient portability. This research introduces dynamic parallel processing to HEAT using OpenMP.

In this research, calibration is defined as a method of obtaining a formula using ground test data. Calibration process is categorized into two parts; calculation and data retrieval. Suko+(2020) implements parallel processing in only calculation part. So, this research introduces dynamic parallel processing in both data retrieval part and calculation one respectively. And this research compares the combinations of the number of threads in nested parallelization. We’re going to demonstrate the fastest solution via zoom in electric poster.