GraphBLAS-like API Design in Functional Style

Kirill Garbar; Igor Erin; Artyom Chernikov; Dmitriy Panfilyonok; Semyon Grigorev

doi:10.32603/2071-2340-2026-1-105-114

Kirill Garbar YandexTechnologies LLC, Lev Tolstoy ul. 16, Moscow, 119021, Russian Federation
Igor Erin YADRO Centr Programmnykh Razrabotok LLC, Rochdelskaya ul. 15, str. 15, Moscow, 123376, Russian Federation
Artyom Chernikov Saint Petersburg State University, 7-9 Universitetskaya Embankment, St. Petersburg, 199034, Russian Federation
Dmitriy Panfilyonok Saint Petersburg State University, 7-9 Universitetskaya Embankment, St. Petersburg, 199034, Russian Federation
Semyon Grigorev Saint Petersburg State University, Universitetskaya nab. 7-9, St. Petersburg, 199034, Russian Federation http://orcid.org/0000-0002-7966-0698

DOI: https://doi.org/10.32603/2071-2340-2026-1-105-114

Keywords: graph analysis, sparse linear algebra, GraphBLAS API, GPGPU, parallel programming, functional programming, .NET, OpenCL, FSharp

Abstract

GraphBLAS API standard describes linear algebra based blocks to build parallel graph analysis algorithms. While it is a promising way to high-performance graph analysis, there are a number of drawbacks such as complicated API, hardness of implementation for GPGPU, and explicit zeroes problem. We show that the utilization of techniques from functional programming can help to solve some GraphBLAS design problems.

Author Biographies

Kirill Garbar, YandexTechnologies LLC, Lev Tolstoy ul. 16, Moscow, 119021, Russian Federation

Backend Developer at Yandex.Technologies LLC, kirillgarbar322@gmail.com

Igor Erin, YADRO Centr Programmnykh Razrabotok LLC, Rochdelskaya ul. 15, str. 15, Moscow, 123376, Russian Federation

младший инженер по разработке ПО в ООО «Ядро Центр Программных Разработок», igor.erin.a@gmail.com

Artyom Chernikov, Saint Petersburg State University, 7-9 Universitetskaya Embankment, St. Petersburg, 199034, Russian Federation

Student at Saint Petersburg State University, artem_chernikov00@list.ru

Dmitriy Panfilyonok, Saint Petersburg State University, 7-9 Universitetskaya Embankment, St. Petersburg, 199034, Russian Federation

Student at Saint Petersburg State University, dmitriy.panfilyonok@gmail.com

Semyon Grigorev, Saint Petersburg State University, Universitetskaya nab. 7-9, St. Petersburg, 199034, Russian Federation

Associate Professor at St. Petersburg State University, s.v.grigoriev@spbu.ru

References

J. Kepner et al., "Mathematical foundations of the graphblas," in 2016 IEEE High Perform. Extreme Comput. Conf. (HPEC), Waltham, MA, USA, 2016, pp. 1-9; doi:10.1109/HPEC.2016.7761646

O. Selvitopi et al., "Distributed many-to-many protein sequence alignment using sparse matrices," in Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal. (SC '20), Atlanta, GA, USA, 2020, pp. 1-14; doi:10.1109/SC41405.2020.000792020

J. Kepner et al., "Enabling massive deep neural networks with the graphblas," in 2017 IEEE High Perform. Extreme Comput. Conf. (HPEC), Waltham, MA, USA, 2017, pp. 1-10; doi:10.1109/HPEC.2017.8091098

T. A. Davis, "Algorithm 1000: Suitesparse: graphblas: Graph algorithms in the language of sparse linear algebra," ACM Trans. Math. Softw., vol. 45, no. 4, pp. 1-25, 2019; doi:10.1145/3322125

A. Buluc and J. R. Gilbert, "The combinatorial blas: Design, implementation, and applications," Int. J. High Perform. Comput. Appl., vol. 25, no. 4, pp. 496-509, 2011; doi:10.1177/1094342011403516

C. Yang, A. Buluc, and J. D. Owens, "Graphblast: A high-performance linear algebra-based graph framework on the gpu," ACM Trans. Math. Softw., vol. 48, no. 1, pp. 1-51, 2022; doi:10.1145/3466795

T. Henriksen et al., "Futhark: Purely functional gpu-programming with nested parallelism and in-place array updates," in Proc. 38th ACM SIGPLAN Conf. Program. Lang. Design Implement. (PLDI 2017), Barcelona, Spain, 2017, pp. 556-571; doi:10.1145/3062341.3062354

T. L. McDonell, M. M. Chakravarty, G. Keller, and B. Lippmeier, "Optimising purely functional gpu programs," SIGPLAN Not., vol. 48, no. 9, pp. 49-60, 2013; doi:10.1145/2544174.2500595

R. Leißa et al., "Anydsl: A partial evaluation framework for programming high-performance libraries," Proc. ACM Program. Lang., vol. 2, no. OOPSLA, 2018; doi:10.1145/3276489