DCSE summerschool 2023 slides and exercise material

This repository contains slides and lab assignments for the summerschool, it will be available for participants during the event.

Overview of the lab sessions

Lab 1, Monday morning: dense linear algebra and data dependencies

We will implement a dense LU decomposition using a block algorithm and OpenMP tasks
details

GPU lab, Monday afternoon: dense matrix-vector multiplication

We will implement General (dense) Matrix-Vector (GEMV) multiplication using OpenMP, and offload the calculation to the GPU. Different CUDA implementations highlight some aspects of GPU hardware-aware programming.
details

Lab 2, Tuesday: Sparse solvers and preconditioners in Trilinos

We will experiment with iterative solvers and preconditinoers from Trilinos, learn how to comile and run applications for multi-core CPUs and GPUs, and investigate performance.
details

Lab 3, Wednesday: performance of SpMV in Trilinos

We will benchmark the Sparse Matrix-Vector (SpMV) multiplication in Trilinos for different matrices, a key operation for iterative linear and eigenvalue solvers.
details

Lab 4, Thursday: communication-avoiding Krylov methods

We will compare standard GMRES, GMRES with polynomial preconditioning and s-step GMRES methods in terms of convergence and performance.
details

Quantum Computing Lab, Friday: Introduction to Quantum Annealing

We will set up and solve a combinatorial optimiation problem on the D-Waves quantum annealer. details