During the first year of master’s studies, I attended an HPC course and completed two major assignments: one utilizing CUDA and the other using MPI.
Matching in genome variant databases Link to heading
The goal was to implement a system for matching genomic variants from a file agains a database (dbNSFP). The solution consists of two phases: database indexing and matching, with the matching phase optimized using CUDA.
For a detailed explanation of my approach, refer to the report. A full description of the task can be found at the link below.
Communication-Efficient Sparse Matrix Multiplication Link to heading
As the title says, the objective was to implement parallel algorithms for multiplying two sparse martices, as described in this paper.
For a detailed explanation of my approach, refer to the report. A full description of the task can be found at the link below.