Many scientific programmers write all of their own code. But many common programming operations amount to linear algebra operations (e.g. matrix multiplication or factorization). Linear algebra libraries such as MKL or ATLAS provide linear algebra implementations optimized for given CPU architectures and the promise of code that is faster to both write and execute.
This session will provide an overview of linear algebra libraries, the routines they provide, and how to use them in code. An overview of higher-level packages and solvers, such as PETSc and Trilinos, will also be provided.