Implementation of Dagger.jl distributed array and MPI features focused on simulation integration

• Grant number: 23/16096-9
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: February 01, 2024
• End date: March 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Leonardo Andre Ambrosio
• Grantee: Felipe de Alcântara Tomé
• Supervisor: Julian Samaroo
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Institution abroad: Massachusetts Institute of Technology (MIT), United States
• Associated to the scholarship: 22/07810-7 - Julia algorithm for simulation of non-diffracting acoustic waves from partial wave expansions and beam shape coefficients, BP.IC

Abstract

This project aims to further explore the ongoing collaboration between the candidate and the MIT advisors on developing a set of tools for distributed array manipulation exploring the Message Passing Interface (MPI) back-end of the Dagger.jl package included in the Julia programming language. The internship period will focus on expanding and completing the MPI sub-package of Dagger.jl and it's linear algebra aspects using the Basic Linear Algebra Sub-routines(BLAS), as well as integrate it with the simulations that were already developed by the candidate, focusing on the parallel calculation of Beam Shape Coefficients and acoustic Frozen Waves. This aims to bridge development and use case scenarios, in order to improve the user experience of Dagger.jl, to test correctness of it's routines and to optimize physical simulations in acoustics making use of the vast expertise of the advisors on building such experiments. The results of this collaboration will be exposed on a scientific report and may result in publications or presentations inside or outside the MIT scientific community.