Incomplete section: Cloud and HPC platforms

[robmoss]

The following chapters are incomplete or missing:

• Available platforms
• Synchronising code and outputs
• Making one thing run faster
• Running many things at the same time

[EamonConway]

I can definitely contribute to a slurm or PBS section. These are definitely the most common workload managers I'm aware of. We should probably also include Nectar?

[robmoss]

The platforms I'm aware of/have experience with are Spartan (slurm), Nectar, and the Melbourne Research Cloud. It's definitely worth including Nectar. What's PBS?

[EamonConway]

It’s a different workload manager to slurm, in my experience the syntax is similar to slurm.

[EamonConway]

I wonder if finding someone with aws experience is worth it as well? I know some bioinformatics pipelines are hosted on aws.

[robmoss]

That's a good question. I suppose it partly depends on what resources are available at each SPARK/SPECTRUM node. Maybe Sebastian Duchene or someone else at the Doherty is worth asking? David Price might know who to ask at the Doherty. I don't know anything about bioinformatics.