# How (if possible) do I allocate more GPU utilization to Ray?

**URL:** <https://discuss.ray.io/t/how-if-possible-do-i-allocate-more-gpu-utilization-to-ray/7551>\
**Category:** Uncategorized\
**Created:** [September 14, 2022, 3:11am UTC](https://discuss.ray.io/t/how-if-possible-do-i-allocate-more-gpu-utilization-to-ray/7551 "2022-09-14T03:11:33Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![aydinguy](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/aydinguy/32/3195_2.png) [@aydinguy](https://discuss.ray.io/u/aydinguy)\
**Post date:** [September 14, 2022, 3:11am UTC](https://discuss.ray.io/t/how-if-possible-do-i-allocate-more-gpu-utilization-to-ray/7551/1 "2022-09-14T03:11:33Z")

</div>

Hello,

I’m a beginner, I’m using Ray rllib to learn a PPO control policy. I’m making repeated calls to the Trainable.train() function.

I’m using the default config, and my GPU utilization is at about 20%. Is it possible to allocate more GPU memory to Ray to speed up training?

Thanks!