High-Productivity GPU Programming Languages

2022 Fiscal Year Final Research Report

• Project/Area Number: 18H03219
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Basic Section 60050:Software-related
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: Masuhara Hidehiko 東京工業大学, 情報理工学院, 教授 (40280937)
• Project Period (FY): 2018-04-01 – 2023-03-31
• Keywords: GPGPU / オブジェクト指向プログラミング / 動的メモリ割り当て

Outline of Final Research Achievements

Towards highly-productive GPGPU programming languages, this project obtained the following results. It proposed a novel algorithm for dynamic memory management allocation, which is one of the primary obstacles that prevents from using object-oriented programming on GPU. The algorithm was implemented as a domain-specific language for C++, called DynaSOAr, which allocates memory as fast as existing memory allocaters, yet allocates more densely, resulting in faster overall application execution. Based on this work, the project further studied on advanced object-oriented language features including inheritance, modularity, and dynamic compilation.

Free Research Field

プログラミング言語

Academic Significance and Societal Importance of the Research Achievements

GPUを搭載した計算機環境がますます一般的になる中で、GPGPUアプリケーションプログラムの記述を容易にすることは研究開始当初から継続して求められており、今後もますます必要となると考えられる。そのような状況下で、これまで性能上の理由から事実上で行えなかった動的メモリ割当ての効率化に成功したことで、GPU上でのオブジェクト指向プログラミングの応用範囲が大きく広がる可能性を示した意義は大きいと考える。