blue-contactMany of the most common underlying algorithms are already performance limited on current tera and petascale platforms. These limitations will only increase, and in some cases become untenable, at exascale. For example, achieving a reasonable computation/communication balance without a major increase in problem size, coping with increasing memory latencies, avoiding global synchronisation points and load imbalance are all important considerations. In addition, exploiting the complex heterogeneous nature of the underlying architecture will become key.

CRESTA evaluated and quantified the limitations of key numerical libraries for the co-design applications and consider whether completely new approaches were required. Achieving scalability to millions of processors is extremely challenging and in many cases it will be necessary to consider alternative algorithms to achieve exascale performance. Where necessary, CRESTA investigated alternative approaches and radical developments to key algorithms to make significant advances in these algorithms for extreme scale computing. In particular, CRESTA focused on linear solvers, particle approaches, FEM and finite volumes and spectral methods.