0

What is the best way to set inputs in CaseIteratorDriver? An example shows direct assignment via: self.driver.case_inputs.param_name = [0,1,2] That doesn't seem to properly get handled when running CaseIteratorDriver in parallel. I tried setting them via self.driver.setsinputs(CaseArray({'x':[0,1,2]}) but that fails too. It appears CaseArray doesn't generate the 'generator' object that set_inputs is expecting. What classes do generate the correct object?

Kitsune4px
  • 5
  • 2

1 Answers1

0

Setting it to a list should have worked fine. What is the behavior you're seeing that makes you think it is not working in parallel?

How are you running it in parallel? By setting sequential=False?

Justin Gray
  • 5,605
  • 1
  • 11
  • 16
  • Checked the openmdao.log file when `sequential=False` and it does look like it's launching the workflow in parallel when I set it with a list, but I don't get any evidence in the output. `eval_responses()` just returns what I think is the last value. How do I get a list of responses that corresponds to the input list? – Kitsune4px May 22 '15 at 22:56
  • 1
    Okay, never mind the class documentation. The results of a set of runs appears to be stored in a variable tree `CaseIteratorDriver.case_outputs` Now that I can find the output, I can see that it is running in parallel. – Kitsune4px Jun 03 '15 at 22:22