# Potential NVSHMEM allocated memory performance issue

**URL:** <https://forums.developer.nvidia.com/t/potential-nvshmem-allocated-memory-performance-issue/275416>\
**Category:** GPU-Accelerated Libraries\
**Tags:** nvshmem\
**Created:** [December 8, 2023, 8:11am UTC](https://forums.developer.nvidia.com/t/potential-nvshmem-allocated-memory-performance-issue/275416 "2023-12-08T08:11:21Z")\
**Posts on this page:** 1\
**Showing post:** 4

<div class="post-metadata">

**Author:** ![jdinan](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/jdinan/32/26862_2.png) [@jdinan](https://forums.developer.nvidia.com/u/jdinan)\
**Post date:** [December 11, 2023, 4:04pm UTC](https://forums.developer.nvidia.com/t/potential-nvshmem-allocated-memory-performance-issue/275416/4 "2023-12-11T16:04:38Z")

</div>

Could you please try setting the environment variable NVSHMEM\_DISABLE\_CUDA\_VMM=1? This disables automatic symmetric heap sizing, so you may also need to increase NVSHMEM\_SYMMETRIC\_SIZE. More information on these env vars is available here: [Environment Variables — NVSHMEM 2.10.1 documentation](https://docs.nvidia.com/nvshmem/api/gen/env.html)

---

_[View the full topic](https://forums.developer.nvidia.com/t/potential-nvshmem-allocated-memory-performance-issue/275416)._