M2 course for 2017-2018: Scheduling at scale
Yves Robert
This course addresses scheduling problems when computing at
extreme scale.
We select a few problems based upon their importance
in real applications. Well, almost :-)
Instead, we select a few problems
because they are beautiful and insightful.
And believe it or not,
some turn out relevant in practice!
Scheduling
- Background
- A few famous approximation algorithms
Failure detection
- Pull and push techniques for process observation
- Randomized protoocols
- Deterministic protoocols
Resilience
- Checkpointing algorothm
- Replication
- Silent errors: detection and correction
Linear algebra algorithms
- Dense factorizations
- Sparse iterative solvers
Scheduling workflows on HPC and cloud platforms