# HyperTuning.jl with multiple GPUs

**URL:** <https://discourse.julialang.org/t/hypertuning-jl-with-multiple-gpus/116721>\
**Category:** General Usage\
**Tags:** flux, hypertuning\
**Created:** [July 7, 2024, 1:44pm UTC](https://discourse.julialang.org/t/hypertuning-jl-with-multiple-gpus/116721 "2024-07-07T13:44:00Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![cirobr](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/cirobr/32/219994_2.png) [@cirobr](https://discourse.julialang.org/u/cirobr)\
**Post date:** [July 7, 2024, 1:44pm UTC](https://discourse.julialang.org/t/hypertuning-jl-with-multiple-gpus/116721/1 "2024-07-07T13:44:00Z")

</div>

Cheers,

I’m working with a machine with two GPUs. The search execution process in parallel for the package is pretty much straightforward and runs smoothly, except for the fact that multiple threads are executed on a single GPU, whilst the other one remains idle.

I have also tried to explicitly create an environment with one process/GPU as in [this link](https://cuda.juliagpu.org/stable/usage/multigpu/#Scenario-1:-One-GPU-per-process). Even so, multiple processes ran on a single GPU.

For using this package, is there a way to allocate one GPU per parallel execution?

If negative, would appreciate suggestions for hyperparameter search with alternative packages.

Thanks in advance.