TOMIOKA Gaku, a first-year master’s student (YOSHIDA Akimasa Laboratory), received the FIT Encouragement Award at the 23rd Information Science and Technology Forum of Information Processing Society of Japan FIT2024 held from September 4 to 6, 2024.
This award is given to the best presentation in each session of the FIT general presentations, chosen by the Chair at their discretion.
TOMIOKA gave a presentation titled “CUDA implementation that takes into account the data layout in explicit particle methods.” This study focused on the explicit particle method used in fluid simulations, and proposed a data layout optimization method for parallel processing on GPUs. This method reduces memory access time by dynamically rearranging the spatial data of particles stored in buckets in AoS (Array of Structure) format. The effectiveness of the proposed method was confirmed by evaluating performance using a 3D water column collapse on a Xeon server equipped with an NVIDIA Tesla K80.
