Enable asynchronous execution and status check

[iazehaf]

Description of the feature

As a developer, I want to start MAPDL asynchronously, so that I can execute some code after the solver start. I want to be able to check the status of the executed analysis at anytime using a get_state method.

One can take as example PyOptiSLang SDK that allows to start an analysis synchronously or asynchronously using the wait_for_finished flag and retrieve the job status using the get_status method.

Steps for implementing the feature

# Just an example
mapdl = launch_mapdl()
mapdl.solution()
mapdl.antype("STATIC")
mapdl.solve(wait_for_finished=False) # Async execution. Sync if wait_for_finished is set to True.

status = mapdl.get_status()

status can take: Running, Completed, Failed.

Useful links and references

No response

[germa89]

Hi @iazehaf

Because of MAPDL server side limitations, it is not possible to get this kind of behaviour unless there are some changes on MAPDL.

Currently, all the gRPC calls are "blocking", meaning they have be to executed one after the other, hence we cannot run an MAPDL command and at the same time, execute some queries on the DB.

We could however "queue" all the gRPC calls, and keep executing other python calls. However those python calls cannot require MAPDL methods.

For instance, assuming async behaviour already implemented.

Blocking call:

mapdl.solution()
mapdl.solve()  # This call will block the rest of the calls since they depends on the MAPDL program
mapdl.post1()
mapdl.post_processing.nodal_displacements()

Non-blocking

import numpy as np
import numpy.linalg as LA

mapdl.solution()
mapdl.solve()

A = np.array([[1, 3, 4], [5, 6, 9], [1, 2, 3]])
b = np.array([1, 2, 1])
x = LA.solve(A, b) # non-blocking up to here

mapdl.parameters['displacements'] = x  # this will be waiting for previous `mapdl.solve` being executed