# \#mpi

**URL:** https://discourse.julialang.org/tag/mpi/170.md

[Latest](https://discourse.julialang.org/latest.md) · [Categories](https://discourse.julialang.org/categories.md) · [Tags](https://discourse.julialang.org/tags.md)

---

## [Developing MPI code using VSCode](https://discourse.julialang.org/t/developing-mpi-code-using-vscode/130108)

<div class="topic-metadata">

**Author:** [@Salmon](https://discourse.julialang.org/u/Salmon)\
**Replies:** 5\
**Last updated:** [December 24, 2025, 1:13am UTC](https://discourse.julialang.org/t/developing-mpi-code-using-vscode/130108 "2025-12-24T01:13:02Z")

</div>

Hi everyone, I want to parallelize some code on multiple nodes and am considering MPI.jl One thing that feels a bit suboptimal, is that MPI seems quite incompatible with the interactive workflow. Ideally, I would like …

---

## [\[San Francisco, CA\] Performance Engineer - GPU Atmospheric Modeling](https://discourse.julialang.org/t/san-francisco-ca-performance-engineer-gpu-atmospheric-modeling/133858)

<div class="topic-metadata">

**Author:** [@giordano](https://discourse.julialang.org/u/giordano)\
**Replies:** 2\
**Last updated:** [November 14, 2025, 3:48am UTC](https://discourse.julialang.org/t/san-francisco-ca-performance-engineer-gpu-atmospheric-modeling/133858 "2025-11-14T03:48:00Z")

</div>

Aeolus Labs are looking for a performance engineer building a next-generation, differentiable atmospheric model for hurricane science and weather forecasting. Model will be based on Oceananigans.jl, experience with the …

---

## [Productive Scalable Distributed Task Scheduling Using an MPI-based Backend for Dagger](https://discourse.julialang.org/t/productive-scalable-distributed-task-scheduling-using-an-mpi-based-backend-for-dagger/131995)

<div class="topic-metadata">

**Author:** [@yanzin00](https://discourse.julialang.org/u/yanzin00)\
**Replies:** 4\
**Last updated:** [September 8, 2025, 4:45pm UTC](https://discourse.julialang.org/t/productive-scalable-distributed-task-scheduling-using-an-mpi-based-backend-for-dagger/131995 "2025-09-08T16:45:56Z")

</div>

Hello Julia Community, especially Dagger and HPC developers. I hope you are doing well this summer and have achieved your goals! I’m here to share: Dagger’s MPI Implementation status My Google Summer of Code at Julia c…

---

## [MPI library mismatch erro HDF5, Netcdf with Oceananigans](https://discourse.julialang.org/t/mpi-library-mismatch-erro-hdf5-netcdf-with-oceananigans/131579)

<div class="topic-metadata">

**Author:** [@farolap](https://discourse.julialang.org/u/farolap)\
**Replies:** 2\
**Last updated:** [September 1, 2025, 6:17am UTC](https://discourse.julialang.org/t/mpi-library-mismatch-erro-hdf5-netcdf-with-oceananigans/131579 "2025-09-01T06:17:17Z")

</div>

Dear all, Im having strong issues with a specific package in Julia while trying to run it in MPI. Initially I define the following libraries: export LD\_LIBRARY\_PATH=/sw/spack-levante/netcdf-c-4.8.1-7dq6g2/lib/:$LD\_LIBR…

---

## [Faster MPI Integration in Dagger](https://discourse.julialang.org/t/faster-mpi-integration-in-dagger/130784)

<div class="topic-metadata">

**Author:** [@yanzin00](https://discourse.julialang.org/u/yanzin00)\
**Replies:** 9\
**Last updated:** [July 27, 2025, 7:39pm UTC](https://discourse.julialang.org/t/faster-mpi-integration-in-dagger/130784 "2025-07-27T19:39:21Z")

</div>

Hello everyone! I hope you’re all doing well as you continue developing your Julia code. I am writing this post to share my progress with the Google Summer of Code project, specifically regarding the MPI integration in D…

---

## [Julia SLURM + BLAS + Multithreading, threads not mapping well leading to poor performance](https://discourse.julialang.org/t/julia-slurm-blas-multithreading-threads-not-mapping-well-leading-to-poor-performance/130188)

<div class="topic-metadata">

**Author:** [@jacksonjhayes](https://discourse.julialang.org/u/jacksonjhayes)\
**Replies:** 5\
**Last updated:** [June 25, 2025, 5:04pm UTC](https://discourse.julialang.org/t/julia-slurm-blas-multithreading-threads-not-mapping-well-leading-to-poor-performance/130188 "2025-06-25T17:04:36Z")

</div>

I am running benchmarking of a code that uses: multithreading BLAS (multithreaded) BLAS (single threaded in parallel on Julia Threads) on NERSC Perlmutter (AMD 2x \[AMD EPYC 7763\] 64 cores x 2 hyperthreads per core) …

---

## [PETSC.jl MPI wrapper](https://discourse.julialang.org/t/petsc-jl-mpi-wrapper/77900)

<div class="topic-metadata">

**Author:** [@Srikumar](https://discourse.julialang.org/u/Srikumar)\
**Replies:** 8\
**Last updated:** [June 7, 2025, 3:23am UTC](https://discourse.julialang.org/t/petsc-jl-mpi-wrapper/77900 "2025-06-07T03:23:08Z")

</div>

Hello, I would love to know if someone has used PETSC wrapper in Julia for MPI? I am planning to develop a parallel code using the PETSC wrapper and would like to know how extensive the PETSC functionality is in Julia.…

---

## [Speeding Up Julia Startup with MPI: Using Scratch Space to Mitigate HDD Access Delay](https://discourse.julialang.org/t/speeding-up-julia-startup-with-mpi-using-scratch-space-to-mitigate-hdd-access-delay/128971)

<div class="topic-metadata">

**Author:** [@0samuraiE](https://discourse.julialang.org/u/0samuraiE)\
**Replies:** 9\
**Last updated:** [June 5, 2025, 1:36pm UTC](https://discourse.julialang.org/t/speeding-up-julia-startup-with-mpi-using-scratch-space-to-mitigate-hdd-access-delay/128971 "2025-06-05T13:36:11Z")

</div>

Hello everyone, I’m currently experiencing a significant slowdown when launching Julia with MPI on my cluster. I suspect that the issue might be due to the long HDD access times. To mitigate this, I am considering copy…

---

## [Pointers of Float to be used in MPI?](https://discourse.julialang.org/t/pointers-of-float-to-be-used-in-mpi/128146)

<div class="topic-metadata">

**Author:** [@jsjie](https://discourse.julialang.org/u/jsjie)\
**Replies:** 0\
**Last updated:** [April 17, 2025, 2:46am UTC](https://discourse.julialang.org/t/pointers-of-float-to-be-used-in-mpi/128146 "2025-04-17T02:46:04Z")

</div>

The following is a MWE that fails: using MPI MPI.Init() mutable struct Foo bar::Float64 end rank = MPI.Comm\_rank(MPI.COMM\_WORLD) foo = Foo(rank) println("$(rank)\\t$(foo.bar)") if rank == 0 MPI.API.MPI\_Recv(Re…

---

## [Distributed.jl vs MPI.jl](https://discourse.julialang.org/t/distributed-jl-vs-mpi-jl/71315)

<div class="topic-metadata">

**Author:** [@Niceno](https://discourse.julialang.org/u/Niceno)\
**Replies:** 26\
**Last updated:** [January 31, 2022, 4:13am UTC](https://discourse.julialang.org/t/distributed-jl-vs-mpi-jl/71315 "2022-01-31T04:13:13Z")

</div>

Hi all, I am relatively new to Julia, but have ported one of my older programs from Fortran to Julia. I picked a program from my collection which doesn’t have any parallelims in it with the intention to implement it fr…

---

## [Distributed Data Parallel training with 2 GPUs fails with Flux.jl on AMD GPUs](https://discourse.julialang.org/t/distributed-data-parallel-training-with-2-gpus-fails-with-flux-jl-on-amd-gpus/125993)

<div class="topic-metadata">

**Author:** [@Alexander-Barth](https://discourse.julialang.org/u/Alexander-Barth)\
**Replies:** 7\
**Last updated:** [February 18, 2025, 8:41am UTC](https://discourse.julialang.org/t/distributed-data-parallel-training-with-2-gpus-fails-with-flux-jl-on-amd-gpus/125993 "2025-02-18T08:41:29Z")

</div>

I am using Flux (with this PR Fix missing imports in FluxMPIExt by Alexander-Barth · Pull Request #2589 · FluxML/Flux.jl · GitHub to fix some MPI related imports) on two AMD GPUs with Distributed Data Parallel (DDP). Ho…

---

## [Save model when training with Distributed Data Parallel](https://discourse.julialang.org/t/save-model-when-training-with-distributed-data-parallel/125790)

<div class="topic-metadata">

**Author:** [@niltsz](https://discourse.julialang.org/u/niltsz)\
**Replies:** 0\
**Last updated:** [February 11, 2025, 1:24pm UTC](https://discourse.julialang.org/t/save-model-when-training-with-distributed-data-parallel/125790 "2025-02-11T13:24:45Z")

</div>

Hey all, I want to parallelize my machine learning model on GPUs using DDP from Flux. I get the code to run with adapting what is written on the Flux GPU Support page, but when I try to save the model I get the error u…

---

## [Why rank 0 cannot receive stdin input?](https://discourse.julialang.org/t/why-rank-0-cannot-receive-stdin-input/124450)

<div class="topic-metadata">

**Author:** [@Blumenkranz](https://discourse.julialang.org/u/Blumenkranz)\
**Replies:** 3\
**Last updated:** [January 6, 2025, 10:22am UTC](https://discourse.julialang.org/t/why-rank-0-cannot-receive-stdin-input/124450 "2025-01-06T10:22:58Z")

</div>

using MPI input\_channel = Channel{String}(1) test = Ref(false) function listen\_for\_input() @async begin println("Rank 0: Listening for input...") while true input = readline(stdin) …

---

## [Nested parallelization in MPI with small outer loop and large inner loop](https://discourse.julialang.org/t/nested-parallelization-in-mpi-with-small-outer-loop-and-large-inner-loop/123610)

<div class="topic-metadata">

**Author:** [@pat-alt](https://discourse.julialang.org/u/pat-alt)\
**Replies:** 2\
**Last updated:** [December 17, 2024, 3:25pm UTC](https://discourse.julialang.org/t/nested-parallelization-in-mpi-with-small-outer-loop-and-large-inner-loop/123610 "2024-12-17T15:25:59Z")

</div>

I’m trying to run a slurm job using MPI, that involves a nested loop, where the outer loop is over a smaller list of size m than the inner loop (size n). For the inner loop, I want to use as many processes as possible np…

---

## [DifferentialEquations.jl+MPI.jl+PencilArrays.jl: Lack of scaling observed](https://discourse.julialang.org/t/differentialequations-jl-mpi-jl-pencilarrays-jl-lack-of-scaling-observed/122998)

<div class="topic-metadata">

**Author:** [@mmesiti](https://discourse.julialang.org/u/mmesiti)\
**Replies:** 20\
**Last updated:** [December 13, 2024, 2:43pm UTC](https://discourse.julialang.org/t/differentialequations-jl-mpi-jl-pencilarrays-jl-lack-of-scaling-observed/122998 "2024-12-13T14:43:54Z")

</div>

I am working on a code that use DifferentialEquations (more precisely, just OrdinaryDiffEqLowOrderRK - thanks for splitting that) and I have done some work to parallelize the computation of the derivative on multiple nod…

---

## [\[ANN\] Vahana.jl - Framework for large-scale agent-based models](https://discourse.julialang.org/t/ann-vahana-jl-framework-for-large-scale-agent-based-models/102024)

<div class="topic-metadata">

**Author:** [@sfuerst](https://discourse.julialang.org/u/sfuerst)\
**Replies:** 19\
**Last updated:** [November 18, 2024, 7:46pm UTC](https://discourse.julialang.org/t/ann-vahana-jl-framework-for-large-scale-agent-based-models/102024 "2024-11-18T19:46:36Z")

</div>

I’m happy to announce the release of Vahana.jl, a framework for (not only) large-scale agent-based models with a focus on complex social networks. There will also be a prerecorded talk at JuliaCon tomorrow. The reposit…

---

## [HDF5 error: MPIExt cannot find libmpi\_usempif08.40.dylib](https://discourse.julialang.org/t/hdf5-error-mpiext-cannot-find-libmpi-usempif08-40-dylib/115900)

<div class="topic-metadata">

**Author:** [@PetrKryslUCSD](https://discourse.julialang.org/u/PetrKryslUCSD)\
**Replies:** 8\
**Last updated:** [November 9, 2024, 8:27pm UTC](https://discourse.julialang.org/t/hdf5-error-mpiext-cannot-find-libmpi-usempif08-40-dylib/115900 "2024-11-09T20:27:26Z")

</div>

Missing /.julia/artifacts/69e85a392db842d919cf5759941f8f3e8a0e569e/lib/libhdf5\_fortran.so" libmpi\_usempif08.so.40: cannot open shared object file: No such file or directory ? Any idea what that is about?

---

## [@profile slows down MPI?](https://discourse.julialang.org/t/profile-slows-down-mpi/121700)

<div class="topic-metadata">

**Author:** [@johnomotani](https://discourse.julialang.org/u/johnomotani)\
**Replies:** 0\
**Last updated:** [October 24, 2024, 2:10pm UTC](https://discourse.julialang.org/t/profile-slows-down-mpi/121700 "2024-10-24T14:10:33Z")

</div>

Is anybody else having problems when trying to @profile an MPI program? I seem to find that when I add @profile my code slows down a lot, with a load of time being spent in MPI calls (a lot more than I would expect). I t…

---

## [Error returned from CUDA function in CUDA-aware MPI multi-GPU test](https://discourse.julialang.org/t/error-returned-from-cuda-function-in-cuda-aware-mpi-multi-gpu-test/121331)

<div class="topic-metadata">

**Author:** [@luciano-drozda](https://discourse.julialang.org/u/luciano-drozda)\
**Replies:** 1\
**Last updated:** [October 23, 2024, 12:59pm UTC](https://discourse.julialang.org/t/error-returned-from-cuda-function-in-cuda-aware-mpi-multi-gpu-test/121331 "2024-10-23T12:59:45Z")

</div>

Hi! I’m coming from this issue here, which seems to not relate to CUDA.jl specifically. Do you think it comes from my openmpi, cuda drivers installations ? Otherwise, could it come from MPI.jl ?

---

## [How do I stop Julia optimising out a variable? (or how to use \`MPI.Ibarrier()\`?)](https://discourse.julialang.org/t/how-do-i-stop-julia-optimising-out-a-variable-or-how-to-use-mpi-ibarrier/121579)

<div class="topic-metadata">

**Author:** [@johnomotani](https://discourse.julialang.org/u/johnomotani)\
**Replies:** 2\
**Last updated:** [October 22, 2024, 1:37pm UTC](https://discourse.julialang.org/t/how-do-i-stop-julia-optimising-out-a-variable-or-how-to-use-mpi-ibarrier/121579 "2024-10-22T13:37:48Z")

</div>

I want to use MPI.Ibarrier() to synchronize some processes where one needs to complete some work before the others can start, so that the first process does not need to wait at the barrier. Pseudo-code: using MPI comm …

---

## [How can I allocate a buffer contains different types of data (for MPI communication, for example)?](https://discourse.julialang.org/t/how-can-i-allocate-a-buffer-contains-different-types-of-data-for-mpi-communication-for-example/121478)

<div class="topic-metadata">

**Author:** [@Blumenkranz](https://discourse.julialang.org/u/Blumenkranz)\
**Replies:** 10\
**Last updated:** [October 20, 2024, 8:12am UTC](https://discourse.julialang.org/t/how-can-i-allocate-a-buffer-contains-different-types-of-data-for-mpi-communication-for-example/121478 "2024-10-20T08:12:23Z")

</div>

I need to communicate data of Vector{Int8} and Vector{Float64} using MPI.jl. How can I achieve this with a single communication?

---

## [Write a String with parallel HDF5](https://discourse.julialang.org/t/write-a-string-with-parallel-hdf5/121025)

<div class="topic-metadata">

**Author:** [@johnomotani](https://discourse.julialang.org/u/johnomotani)\
**Replies:** 1\
**Last updated:** [October 16, 2024, 6:09am UTC](https://discourse.julialang.org/t/write-a-string-with-parallel-hdf5/121025 "2024-10-16T06:09:21Z")

</div>

Does anyone know the ‘correct’ way to write a scalar variable (especially a String) using parallel HDF5? I think I have a way of doing it, but it’s so ugly I’m hoping there’s a better way! The problem: suppose I have a …

---

## [PythonCall.jl with CondaPkg.jl in MPI scenario](https://discourse.julialang.org/t/pythoncall-jl-with-condapkg-jl-in-mpi-scenario/121191)

<div class="topic-metadata">

**Author:** [@J-C-Q](https://discourse.julialang.org/u/J-C-Q)\
**Replies:** 2\
**Last updated:** [October 11, 2024, 12:37pm UTC](https://discourse.julialang.org/t/pythoncall-jl-with-condapkg-jl-in-mpi-scenario/121191 "2024-10-11T12:37:31Z")

</div>

Hi everyone, I’m working on a julia package that has PythonCall.jl and CondaPkg.jl in its dependencies. Now, I want to do some calculations using my package on a cluster with MPI.jl. The problem is (this is my understa…

---

## [Heap-size-hint in MPIClusterManager.jl for out of memory problems](https://discourse.julialang.org/t/heap-size-hint-in-mpiclustermanager-jl-for-out-of-memory-problems/121050)

<div class="topic-metadata">

**Author:** [@sparrowhawk](https://discourse.julialang.org/u/sparrowhawk)\
**Replies:** 3\
**Last updated:** [October 8, 2024, 2:23pm UTC](https://discourse.julialang.org/t/heap-size-hint-in-mpiclustermanager-jl-for-out-of-memory-problems/121050 "2024-10-08T14:23:15Z")

</div>

Hi, for the MPI transport option, where only MPI can be used for communication on the cluster, I’ve gotten the example from MPIClusterManagers.jl working well for my use case. However, I would like to add a heap size hin…

---

## [How to profile Julia MPI code?](https://discourse.julialang.org/t/how-to-profile-julia-mpi-code/57136)

<div class="topic-metadata">

**Author:** [@PolarizedPoutine](https://discourse.julialang.org/u/PolarizedPoutine)\
**Replies:** 17\
**Last updated:** [October 3, 2024, 12:16am UTC](https://discourse.julialang.org/t/how-to-profile-julia-mpi-code/57136 "2024-10-03T00:16:05Z")

</div>

How do people usually profile Julia MPI code to find bottlenecks and reduce communication overhead? As an MPI noob I thought there would be some profiler tools out (kinda like cuprof and NVIDIA Nsight) there but they se…

---

## [Testing parallelization with MPI and Test](https://discourse.julialang.org/t/testing-parallelization-with-mpi-and-test/114040)

<div class="topic-metadata">

**Author:** [@mesonepigreco](https://discourse.julialang.org/u/mesonepigreco)\
**Replies:** 2\
**Last updated:** [September 23, 2024, 4:40am UTC](https://discourse.julialang.org/t/testing-parallelization-with-mpi-and-test/114040 "2024-09-23T04:40:06Z")

</div>

Hello everyone, I am developing a package that implements multiprocessing with MPI.jl for applications in HPC facilities. I have a full test suite for my package, which can be run with the usual “\]” test from the REPL …

---

## [JUHPC: HPC setup for Juliaup, Julia and some HPC key packages](https://discourse.julialang.org/t/juhpc-hpc-setup-for-juliaup-julia-and-some-hpc-key-packages/119556)

<div class="topic-metadata">

**Author:** [@samo](https://discourse.julialang.org/u/samo)\
**Replies:** 0\
**Last updated:** [September 18, 2024, 3:06pm UTC](https://discourse.julialang.org/t/juhpc-hpc-setup-for-juliaup-julia-and-some-hpc-key-packages/119556 "2024-09-18T15:06:19Z")

</div>

@channel Dear Julia HPC community! After an odyssey through stacked environments with pre-installed packages and other unsatisfying solutions, the Julia HPC community is now able to do a good Julia HPC setup. Yet, the …

---

## [Parallelization issues](https://discourse.julialang.org/t/parallelization-issues/119467)

<div class="topic-metadata">

**Author:** [@ditfurth](https://discourse.julialang.org/u/ditfurth)\
**Replies:** 4\
**Last updated:** [September 19, 2024, 12:56pm UTC](https://discourse.julialang.org/t/parallelization-issues/119467 "2024-09-19T12:56:29Z")

</div>

Hi all, I am having parallelization issues. I am trying to utilize multiple computers in an mpi call and have tried a lot of things. So because it is mpi, the memory will not be shared. I can divide my dataset with an i…

---

## [Code that uses MPI.jl does not run on cluster](https://discourse.julialang.org/t/code-that-uses-mpi-jl-does-not-run-on-cluster/118167)

<div class="topic-metadata">

**Author:** [@Uranium238](https://discourse.julialang.org/u/Uranium238)\
**Replies:** 24\
**Last updated:** [August 26, 2024, 3:04pm UTC](https://discourse.julialang.org/t/code-that-uses-mpi-jl-does-not-run-on-cluster/118167 "2024-08-26T15:04:14Z")

</div>

I have written a version of my code that uses MPI.jl to achieve a potentially large scale parallelization. I already had MPI installed on my PC and used the instructions from here to make sure that the same mpiexec is be…

---

## [Bitten by BLAS threading. Again. (Interaction w/ MPI.)](https://discourse.julialang.org/t/bitten-by-blas-threading-again-interaction-w-mpi/118533)

<div class="topic-metadata">

**Author:** [@PetrKryslUCSD](https://discourse.julialang.org/u/PetrKryslUCSD)\
**Replies:** 2\
**Last updated:** [August 24, 2024, 2:19pm UTC](https://discourse.julialang.org/t/bitten-by-blas-threading-again-interaction-w-mpi/118533 "2024-08-24T14:19:10Z")

</div>

I ran some MPI codes. On a Mac 2 Ultra (16 performance cores), and on a Grace+Grace system (two sockets, 72 cores each). The runs were carried out with 4 processes on a single machine (node). There was a weird mismatch …

[Next page](https://discourse.julialang.org/tag/mpi/170.md?match_all_tags=true&page=1&tags%5B%5D=mpi)