# GPU parallelization error, ranks not equal to gpu number

**URL:** <https://pyfr.discourse.group/t/gpu-parallelization-error-ranks-not-equal-to-gpu-number/661>\
**Category:** Errors\
**Tags:** cuda, mpi\
**Created:** [1 August 2022 13:43 UTC](https://pyfr.discourse.group/t/gpu-parallelization-error-ranks-not-equal-to-gpu-number/661 "2022-08-01T13:43:41Z")\
**Posts on this page:** 1\
**Showing post:** 4

<div class="post-metadata">

**Author:** ![WillT](https://avatars.discourse-cdn.com/v4/letter/w/bc79bd/32.png) [@WillT](https://pyfr.discourse.group/u/WillT)\
**Post date:** [1 August 2022 14:27 UTC](https://pyfr.discourse.group/t/gpu-parallelization-error-ranks-not-equal-to-gpu-number/661/4 "2022-08-01T14:27:29Z")

</div>

Did try the solutions from this post when you had a similar problem?

> [@Runtime error about 2d euler vortex example](https://pyfr.discourse.group/t/runtime-error-about-2d-euler-vortex-example/636/4):
>
> Yes, if you run PyFR with a command something like mpiexec -n 4 pyfr run -b cuda ... that should do the trick. If you haven’t already it’s probably worth compiling MPI with the CUDA aware flag, see here for more details: [Performance Tuning — Documentation](https://pyfr.readthedocs.io/en/latest/performance_tuning.html#cuda-backend)

---

_[View the full topic](https://pyfr.discourse.group/t/gpu-parallelization-error-ranks-not-equal-to-gpu-number/661)._