Abstract
This paper reports on the acceleration of a standard, lattice-based numerical algorithm that is widely used in finance for pricing a class of fixed-income vanilla derivatives. We start with a high-level algorithmic specification, exhibiting irregular nested parallelism, which is challenging to map efficiently to GPU hardware. From it we systematically derive and optimize two CUDA implementations, which utilize only the outer or all levels of parallelism, respectively. A detailed evaluation demonstrates (i) the high impact of the proposed optimizations, (ii) the complementary strength and weaknesses of the two GPU versions, and that (iii) they are on average 2.4× faster than our well-tuned CPU-parallel implementation (OpenMP+AVX2) running on 104 hardware threads, and by 3-to-4 order of magnitude faster than an OpenMP-parallel implementation using the popular QuantLib library.
Originalsprog | Engelsk |
---|---|
Titel | ARRAY 2021 - Proceedings of the 7th ACM SIGPLAN International Workshop on Libraries, Languages and Compilers for Array Programming, co-located with PLDI 2021 |
Redaktører | Tze Meng Low, Jeremy Gibbons |
Antal sider | 12 |
Forlag | Association for Computing Machinery, Inc. |
Publikationsdato | 2021 |
Sider | 27-38 |
Artikelnummer | 3464309 |
ISBN (Elektronisk) | 978-1-4503-8466-7 |
DOI | |
Status | Udgivet - 2021 |
Begivenhed | 7th ACM SIGPLAN International Workshop on Libraries, Languages, and Compilers for Array Programming, ARRAY 2021, held in association with PLDI 2021 - Virtual, Online, Canada Varighed: 21 jun. 2021 → … |
Konference
Konference | 7th ACM SIGPLAN International Workshop on Libraries, Languages, and Compilers for Array Programming, ARRAY 2021, held in association with PLDI 2021 |
---|---|
Land/Område | Canada |
By | Virtual, Online |
Periode | 21/06/2021 → … |
Sponsor | ACM SIGPLAN |
Bibliografisk note
Funding Information:This research has been partially supported by the Independent Research Fund Denmark grant under the research project FUTHARK: Functional Technology for High-performance Architectures and by Innovation Fund Denmark under the research project ref.no. 5189-00224B.
Publisher Copyright:
© 2021 ACM.