This project consists of three tasks: 1) analyzing the execution time of chol(A) for Cholesky decomposition to verify cubic complexity. 2) working with sparse matrices stored efficiently using few vectors, similar to the LAPACK format, to solve tridiagonal systems, 3) implementing tensor operations with multidimensional arrays, such as ttt or ttv.

Homogeneous and Heterogeneous Implementations of a tridiagonal solver on Intel® Xeon® E-2176G with oneMKL getrs() routine (Intel® Devmesh Project)