Skip to content

Latest commit

 

History

History
13 lines (10 loc) · 769 Bytes

README.md

File metadata and controls

13 lines (10 loc) · 769 Bytes

spark_in_dakota

Try to tune spark configuration using Sandia labs Dakota

Dakota needs several inputs.

  • an input .in file which is documented here. This defines the optimization algorithms that Dakota will use to search the parameter space.
  • An executable script that takes a parameter file as input and writes to standard out
  • A parameter file to use as an example/template. The fields can be specified to Dakota by position
  • A driver script, which can be a python2.7 file that the Dakota UI can generate using a wizard
  • An example output file, or specification of the fields that are under analysis