Skip to content

ACI REF (not Harvard)

Bob Freeman edited this page Aug 25, 2015 · 15 revisions

So I was talking to Bob about this a bit while at the XSEDE meeting, and right now with Martin. How does this look to consolidate/remove overlap and get a good flow? I would like comments and then perhaps we can do a division of the labor and work these up to finish up the software carpentry training -- Anita

  1. Introduction to HPC – General HPC terms, architecture, file systems, data storage
  2. Accessing the resource – SSH, file transfer, environment, modules
  3. Batch system – job submission, queues, interactive batch jobs, job arrays
  4. Concurrent Serial Strategies
  5. Types of Parallelism and their use
  6. Evaluation of scaling and diagnosing of job failures

What should be in six half-hour lessons?

  1. Accessing the resource. SSH, File Transfers
  2. Introduction to the resource. Job Scheduling, Queues/Partitions, File systems
  3. Understanding Types of Parallelism
  4. Software and MPI. Module usage, MPI jobs
  5. Job Submission. Batch Jobs
  6. Sixth formative assessment

Summative assessment on :

Top Twelve From Software Carpentry Harvard

  • Accessing the resource. SSH, File Transfers - 7
  • Introduction to the resource. Job Scheduling, Queues/Partitions, File systems - 7
  • Interactive Jobs - 2
  • Job Submission. Batch Jobs - 7
  • Job Arrays - 1
  • Software and MPI. Module usage, MPI jobs - 6
  • File Handling - 2
  • Parallel/Concurrent serial strategies/workflows - 3
  • Evaluating Program Scaling - 1
  • Diagnosing Job Failure
  • Understanding Types of Parallelism - 6
  • Parallel Data Processing

Top topics (definite)

  • Introduction to the resource. Job Scheduling, Queues/Partitions, File systems - 7
  • Accessing the resource. SSH, File Transfers - 7
  • Job Submission. Batch Jobs - 7
  • Understanding Types of Parallelism - 6
  • Software and MPI. Module usage, MPI jobs - 6

Last topic (revote):

  • Interactive Jobs -
  • Job Arrays -
  • File Handling -
  • Parallel/Concurrent serial strategies/workflows -
  • Evaluating Program Scaling -
  • Diagnosing Job Failure -
  • Parallel Data Processing -

Clone this wiki locally