quaquel / EMAworkbench / 6683094681

81%

master: 93%

push

github

Implement MPIEvaluator

evaluators.py: Correctly do conditional import

If you only need it in one method of one class, just import it in that one method of that one class (and not in the __init__).

Makes sure mpi4py is only required when the MPIEvaluator is used. Challenging part will be conditionally importing MPI to establish the rank based on which evaluator is used.

[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Update netlogo.py

Update paths used in NetLogo.py for DelftBlue

Update model.py

model.py: Improve ValueError by printing model_file

workbench: Add MPIEvaluator to init

Update DelftBlue commands netlogo.txt

Rewrite for NetLogo.

Current problem: Extracting with tar -xzf NetLogo-6.2.2-64.tgz takes extremely long.

Create DelftBlue commands netlogo.txt

Identical to the current DelftBlue commands.txt, to be able to track changes

[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Scripts: Getting NetLogo working on WSL

6.2.2 works, 6.3.0 give very strange errors pointing to Java ASM incompatibility. Possibly an error in PyNetLogo. See https://github.com/NetLogo/NetLogo/issues/2171

Refactor MPIEvaluator to optimize model handling and runner instantiation

This commit introduces two primary architectural modifications to the MPIEvaluator class, aimed at improving efficiency when running experiments on HPC systems:

1. **Singleton ExperimentRunner**:
    Previously, the `ExperimentRunner` was instantiated for every individual experiment. This approach could introduce unnecessary overhead, especially when dealing with a large number of experiments. Instead, we've adopted a pattern where the `ExperimentRunner` is instantiated once and shared among all the worker processes in the MPI pool. This is achieved using an initializer function `mpi_initializer` which sets up a global `ExperimentRunner` for all the worker processes.

2. **O... (continued)

28 of 28 new or added lines in 1 file covered. (100.0%)

4627 of 5738 relevant lines covered (80.64%)

0.81 hits per line