Informacionnye TehnologiiInformacionnye TehnologiiИнформационные технологии1684-6400New Technologies Publishing House70228510.17587/it.31.322-330Software engineeringПрограммная инженерияReview ArticlePerformance analysis of symbolic algebra libraries in the Julia programming languageАнализ производительности библиотек символьной алгебры языка программирования JuliaZgodaIu. N.ЗгодаЮ. Н.
Russian Federation

Specialist, Education Center of Digital Competence

специалист 2 кат., Образовательный центр цифровых компетенций

yurii.zgoda@mail.ruSaint Petersburg State University of Architecture and Civil EngineeringСанкт-Петербургский государственный архитектурно-строительный университет150620253163223300602202606022026Copyright ©; 2025, Informacionnye TehnologiiCopyright ©; 2025, Информационные технологии2025Informacionnye TehnologiiИнформационные технологииhttps://journals.eco-vector.com/1684-6400/article/view/702285

In many scientific fields not only numerical methods, but also symbolic computations tools are used. However, due to the lack of research in the field of performance of symbolic manipulation libraries, it is often hard to determine the optimal software package or library for high-performance scientific research. The purpose of this work is to perform a comparative analysis of various computer algebra libraries for the Julia programming language in terms of computational performance. As part of this goal, a series of benchmarks were developed and carried using the Symbolics.jl, SymEngine.jl, SymPy.jl libraries. Each benchmark was designed to test different common use cases, such as manipulations with large symbolic expressions, differentiation, symbolic expansion, simplification and substitution. Results of each benchmark were analyzed, and the advantages and disadvantages of different libraries were explored. The source code of benchmarks was discussed, and different performance benchmarking features of these libraries were described. In the process of developing benchmarks, various features of computer algebra libraries were discovered, such as effective algorithms for large sum or product generators. The research showed that in most test cases, the SymEngine.jl library demonstrated the highest performance measures. In cases where SymEngine lacks features, SymPy.jl could be used for additional symbolic expression manipulation operations. The results of this paper could help researchers in different fields of science make an informed choice of software library for their tasks.

Во многих областях науки возникает необходимость решения вычислительных задач не только путем применения численных методов, но и с использованием средств компьютерной алгебры. В данной работе проводится сравнительный анализ трех библиотек символьной алгебры языка программирования Julia: Symbolics.jl, SymEngine.jl и SymPy.jl. Выполнен анализ производительности этих библиотек на различных тестовых задачах, рассмотрены преимущества и недостатки каждой из библиотек.

symbolic computationsJulia programming languageSymbolics.jlSymEngineSymPyfunction compilationhigh-performance computingcomparative analysissymbolic expression generationanalytical solutionсимвольные вычисленияязык программирования JuliaSymbolics.jlSymEngineSymPyкомпиляция функцийвысокопроизводительные вычислениясравнительный анализгенерация символьных выраженийаналитическое решениеGathen J. von zur, Gerhard J. Modern Computer Algebra, United Kingdom, Cambridge University Press, 2003, 804 pp.Gathen J. von zur, Gerhard J. Modern Computer Algebra. United Kingdom: Cambridge University Press, 2003. 804 p.Cohen J. S. Computer Algebra and Symbolic Computation: Mathematical Methods, New York: A K Peters/CRC Press, 2003, 472 pp., doi: 10.1201/9781439863701.Cohen J. S. Computer Algebra and Symbolic Computation: Mathematical Methods. New York: A K Peters/CRC Press, 2003. 472 p. DOI: 10.1201/9781439863701.Navarro A. K. W., Vassiliadis V. S. Computer algebra systems coming of age: Dynamic simulation and optimization of DAE systems in Mathematica™, Computers & Chemical Engineering, 2014, vol. 62, pp. 125-138, doi: 10.1016/j.compchemeng.2013.11.004.Navarro A. K. W., Vassiliadis V. S. Computer algebra systems coming of age: Dynamic simulation and optimization of DAE systems in Mathematica™ // Computers & Chemical Engineering. 2014. Vol. 62. P. 125-138. DOI: 10.1016/j.compchemeng.2013.11.004.Iglesias A. Special section: Computer algebra systems and their applications, CASA'2003-06: Selected papers, Future Generation Computer Systems, 2007, vol. 23, no. 5, pp. 714-715, doi: 10.1016/j.future.2006.10.002.Iglesias A. Special section: Computer algebra systems and their applications, CASA'2003-06: Selected papers // Future Generation Computer Systems. 2007. Vol. 23, N. 5. P. 714-715. DOI: 10.1016/j.future.2006.10.002.?Pavlovic M. N. Symbolic computation in structural engineering, Computers & Structures, 2003, vol. 81, no. 22, pp. 2121-2136, doi: 10.1016/S0045-7949(03)00286-4.Pavlovic M. N. Symbolic computation in structural engineering // Computers & Structures. 2003. Vol. 81, N. 22. P. 2121-2136. DOI: 10.1016/S0045-7949(03)00286-4.Bouzidi Y., Poteaux A., Quadrat A. Computer algebra methods for the stability analysis of differential systems with commensurate time-delays, IFAC-PapersOnLine, 2016, vol. 49, no. 10, pp. 194-199, doi: 10.1016/j.ifacol.2016.07.528.Bouzidi Y., Poteaux A., Quadrat A. Computer algebra methods for the stability analysis of differential systems with commensurate time-delays // IFAC-PapersOnLine. 2016. Vol. 49, N. 10. P. 194-199. DOI: 10.1016/j.ifacol.2016.07.528.Li R., Zhu H., Banach R. An algebraic approach to simulation and verification for cyber-physical systems with shared-variable concurrency, Journal of Logical and Algebraic Methods in Programming, 2024, vol. 139, pp. 100973, doi: 10.1016/j.jlamp.2024.100973.Li R., Zhu H., Banach R. An algebraic approach to simulation and verification for cyber-physical systems with shared-variable concurrency // Journal of Logical and Algebraic Methods in Programming. 2024. Vol. 139. P. 100973. DOI: 10.1016/j.jlamp.2024.100973.Malykh M. D., Sevastianov L. A., Yu Y. On symbolic integration of algebraic functions, Journal of Symbolic Computation, 2021, vol. 104, pp. 563-579, doi: 10.1016/j.jsc.2020.09.002.Malykh M. D., Sevastianov L. A., Yu Y. On symbolic integration of algebraic functions // Journal of Symbolic Computation. 2021. Vol. 104. P. 563-579. DOI: 10.1016/j.jsc.2020.09.002.Szriftgiser P., Cheb-Terrab E. S. Computer algebra in physics: The hidden SO(4) symmetry of the hydrogen atom, Computer Physics Communications, 2021, vol. 268, pp. 108076, doi: 10.1016/j.cpc.2021.108076.Szriftgiser P., Cheb-Terrab E. S. Computer algebra in physics: The hidden SO(4) symmetry of the hydrogen atom // Computer Physics Communications. 2021. Vol. 268. P. 108076. DOI: 10.1016/j.cpc.2021.108076.Wu Q., Li Z. FeAmGen.jl: A Julia program for Feynman Amplitude Generation, Computer Physics Communications, 2024, vol. 301, pp. 109230, doi: 10.1016/j.cpc.2024.109230.Wu Q., Li Z. FeAmGen.jl: A Julia program for Feynman Amplitude Generation // Computer Physics Communications. 2024. Vol. 301. P. 109230. DOI: 10.1016/j.cpc.2024.109230.Karak S. K., Chatterjee S., Bandopadhyay S. Mathematical modelling of the physical and mechanical properties of nano-Y2O3 dispersed ferritic alloys using evolutionary algorithm-based neural network, Powder Technology, 2015, vol. 274, pp. 217-226, doi: 10.1016/j.powtec.2015.01.028.Karak S. K., Chatterjee S., Bandopadhyay S. Mathematical modelling of the physical and mechanical properties of nano-Y2O3 dispersed ferritic alloys using evolutionary algorithm-based neural network // Powder Technology. 2015. Vol. 274. P. 217-226. DOI: 10.1016/j.powtec.2015.01.028.Winkler F. Symbolic computation in algebra, geometry, and differential equations, Information and Computation, 2024, pp. 105200, doi: 10.1016/j.ic.2024.105200.Winkler F. Symbolic computation in algebra, geometry, and differential equations // Information and Computation. 2024. P. 105200. DOI: 10.1016/j.ic.2024.105200.Zgoda Iu. N., Semenov A. A. High performance computation of thin shell constructions with the use of parallel computations and GPUs, Vychislitel'nye Tekhnologii, 2022, vol. 27, no. 6, pp. 45-57, doi: 10.25743/ICT.2022.27.6.005 (in Russian).Згода Ю. Н., Семенов А. А. Высокопроизводительный расчет тонкостенных оболочечных конструкций с использованием параллельных вычислений и графических ускорителей // Вычислительные Технологии. 2022. Т. 27, № 6. С. 45-57. DOI: 10.25743/ICT.2022.27.6.005.Hrdina J., Navrat A., Zalabova L. Symmetries in geometric control theory using Maple, Mathematics and Computers in Simulation, 2021, vol. 190, pp. 474-493, doi: 10.1016/j.matcom.2021.05.034.Hrdina J., Navrat A., Zalabova L. Symmetries in geometric control theory using Maple // Mathematics and Computers in Simulation. 2021. Vol. 190. P. 474-493. DOI: 10.1016/j.mat-com.2021.05.034.Roanes-Lozano E., Galan-Garcla J. L., Aguilera-Venegas G. A computer approach to overtaking station track layout diagram design using graphs. An alternative track diagram proposal for these stations, Journal of Computational and Applied Mathematics, 2021, vol. 391, pp. 113455, doi: 10.1016/j.cam.2021.113455.Roanes-Lozano E., Galan-Garcla J. L., Aguilera-Venegas G. A computer approach to overtaking station track layout diagram design using graphs. An alternative track diagram proposal for these stations // Journal of Computational and Applied Mathematics. 2021. Vol. 391. P. 113455. DOI: 10.1016/j.cam.2021.113455.Anand C. K., Bain A. D., Nie Z. Simulation of steady-state NMR of coupled systems using Liouville space and computer algebra methods Journal of Magnetic Resonance, 2007, vol. 189, no. 2, pp. 200-208, doi: 10.1016/j.jmr.2007.09.012.Anand C. K., Bain A. D., Nie Z. Simulation of steady-state NMR of coupled systems using Liouville space and computer algebra methods // Journal of Magnetic Resonance. 2007. Vol. 189, N. 2. P. 200-208. DOI: 10.1016/jjmr.2007.09.012.Irtegov V. D., Titorenko T. N. Using the system "Mathematica" in problems of mechanics, Mathematics and Computers in Simulation, 2001, vol. 57, no. 3, pp. 227-237, doi: 10.1016/S0378-4754(01)00341-X.Irtegov V. D., Titorenko T. N. Using the system "Mathematica" in problems of mechanics // Mathematics and Computers in Simulation. 2001. Vol. 57, N. 3. P. 227-237. DOI: 10.1016/S0378-4754(01)00341-X.Amata L., Oliveri F., Sgroi E. Optimal Systems of Lie Subalgebras: A Computational Approach, Journal of Geometry and Physics, 2024, pp. 105290, doi: 10.1016/j.geomphys.2024.105290.Amata L., Oliveri F., Sgroi E. Optimal Systems of Lie Subalgebras: A Computational Approach // Journal of Geometry and Physics. 2024. P. 105290. DOI: 10.1016/j.geomphys.2024.105290.Stocco D., Bertolazzi E. Symbolic matrix factorization for differential-algebraic equations index reduction, Journal of Computational and Applied Mathematics, 2024, vol. 448, pp. 115898. doi: 10.1016/j.cam.2024.115898.Stocco D., Bertolazzi E. Symbolic matrix factorization for differential-algebraic equations index reduction // Journal of Computational and Applied Mathematics. 2024. Vol. 448. P. 115898. DOI: 10.1016/j.cam.2024.115898.Meurer A. et al. SymPy: symbolic computing in Python, PeerJ Computer Science. PeerJ Inc, 2017, vol. 3, pp. e103, doi: 10.7717/peerj-cs.103.Meurer A. et al. SymPy: symbolic computing in Python // PeerJ Computer Science. PeerJ Inc., 2017. Vol. 3. P. e103. DOI: 10.7717/peerj-cs.103.Site Rationale - Symengine latest documentation [electronic resource], available at: https://symengine.org/index.html (access date: 03.08.2024).Site Rationale - Symengine latest documentation [электронный ресурс]. URL: https://symengine.org/index.html (дата доступа: 03.08.2024).Bezanson J. et al. Julia: A Fresh Approach to Numerical Computing, SIAM Review. Society for Industrial and Applied Mathematics, 2017, vol. 59, no. 1, pp. 65-98, doi: 10.1137/141000671.Bezanson J. et al. Julia: A Fresh Approach to Numerical Computing // SIAM Review. Society for Industrial and Applied Mathematics, 2017. Vol. 59, N. 1. P. 65-98. DOI: 10.1137/141000671.Gowda S. et al. High-performance symbolic-numerics via multiple dispatch, ACM Communications in Computer Algebra, 2022, vol. 55, no. 3, pp. 92-96, doi: 10.1145/3511528.3511535.Gowda S. et al. High-performance symbolic-numerics via multiple dispatch // ACM Commun. Comput. Algebra. 2022. Vol. 55, N. 3. P. 92-96. DOI: 10.1145/3511528.3511535.symengine/symengine: SymEngine is a fast symbolic manipulation library, written in C++ [electronic resource], available at: https://github.com/symengine/symengine (access date: 03.08.2024).symengine/symengine: SymEngine is a fast symbolic manipulation library, written in C++ [электронный ресурс] URL: https://github.com/symengine/symengine (дата доступа: 03.08.2024).symengine/AUTHORS at master symengine/symengine [electronic resource], GitHub, available at: https://github.com/symengine/symengine/blob/master/AUTHORS (access date: 04.08.2024).symengine/AUTHORS at master symengine/symengine [электронный ресурс] // GitHub. URL: https://github.com/symengine/symengine/blob/master/AUTHORS (дата доступа: 04.08.2024).Zgoda Iu. N., Semenov A. A. Performance analysis of Python symbolic mathematics libraries, Cifrovizaciya i kiberbezopasnost': sovremennaya teoriya i praktika. Sb. nauch. tr. Omsk: SibADI, 2021, pp. 89-92 (In Russian).Згода Ю. Н., Семенов А. А. Анализ производительности библиотек символьной математики для Python // Цифровизация и кибербезопасность: современная теория и практика. Сб. науч. тр. Омск: СибАДИ. 2021. C. 89-92.Chen J., Revels J. Robust benchmarking in noisy environments, arXiv e-prints, 2016, doi: 10.48550/arXiv.1608.04295.Chen J., Revels J. Robust benchmarking in noisy environments // arXiv e-prints. 2016. DOI: 10.48550/arXiv.1608.04295.Akinshin A. Pro.NET Benchmarking: The Art of Performance Measurement, SPb, Piter, 2022, 576 p. (In Russian).Акиньшин А. Профессиональный бенчмарк: искусство измерения производительности. СПб.: Питер, 2022. 576 c.Zgoda Iu.N. Original approach to generation of approximation functions for high-performance shell structures simulation, Vestnik Voronezhskogo gosudarstvennogo tekhnicheskogo universiteta, 2023, vol. 19, no. 5, pp. 8-15, doi: 10.36622/VSTU.2023.19.5.001 (in Russian).Згода Ю. Н. Построение аппроксимационных функций для высокопроизводительного моделирования оболочек // Вестник Воронежского государственного технического университета. 2023. Т. 19, № 5. С. 8-15. DOI: 10.36622/VSTU.2023.19.5.001.How to generate sums effectively via creating an Add object o Issue #67 o jverzani/SymPyCorejl [electronic resource], GitHub, available at: https://github.com/jverzani/SymPyCore.jl/issues/67 (access date: 05.08.2024).How to generate sums effectively via creating an Add object o Issue #67 o jverzani/SymPyCorejl [электронный ресурс] // GitHub. URL: https://github.com/jverzani/SymPyCore.jl/issues/67 (дата доступа: 05.08.2024).