HPC

Sparse-Matrix Vector Multiplication (SpMV)
  • Optimized Sparse-Matrix Vector Multiplication using different parallelism and concurency methods.
  • I developed SpMV method using the following different framework:
    • multi threading
    • OpenMP
    • Rayon-rs
    • Vector processor (intel AVX-2)
    • SIMD (using Cuda on Nvidia GPU)
  • Since each approach takes a different prespective to the problem, a different set of algorithm and data structure was needed to achieve peak performance in each senario. As part of the task, I studied and implemented a variety of solutions proposed in defferent researches and evaluated their performance. (full details in the repo)
Github