# Parallel runs for custom objective function with high computational cost

**URL:** <https://uqworld.org/t/parallel-runs-for-custom-objective-function-with-high-computational-cost/1645>\
**Category:** Community Q&A and How To\
**Created:** [June 22, 2022, 9:38am UTC](https://uqworld.org/t/parallel-runs-for-custom-objective-function-with-high-computational-cost/1645 "2022-06-22T09:38:40Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Omidreza](https://uqworld.org/user_avatar/uqworld.org/omidreza/32/1098_2.png) [@Omidreza](https://uqworld.org/u/Omidreza)\
**Post date:** [June 22, 2022, 9:38am UTC](https://uqworld.org/t/parallel-runs-for-custom-objective-function-with-high-computational-cost/1645/1 "2022-06-22T09:38:40Z")

</div>

Dear Community,

I would like to use the UQ bayesian inversion for a custom likelihood function (objective function) which includes running a simulation which takes a long time (each iteration over the objective function usually takes around 1 minute on a single cpu core).

Is there a possibility in UQ to make the sampling or the inversion process run the objective function in parallel so that I can use all the cpu cores?

something similar to what MATLAB has for its optimisation toolbox as the parallel pool.

Thanks,