Problem A basic test of CUDA-aware MPI with MPI.jl fails on both our Cray supercomputer and on another cluster: 1) On Cray system 1.1) Test and error: omlins@dom101:~> export MPICH_RDMA_ENABLED_CUDA=1 omlins@nid00002:~> julia julia> using MPI julia> using CuArrays julia> MPI.Init() julia> …

I just ran over this old topic and will quickly update it: meanwhile, CUDA-aware MPI works on the Cray system by just setting: export MPICH_RDMA_ENABLED_CUDA=1 export JULIA_CUDA_USE_BINARYBUILDER=false I guess the problem was simply that CUDA.jl did not use the system installation in the past as i…

Error/segfault in basic test of CUDA-aware MPI

Specific Domains Julia at Scale

simonbyrne January 28, 2020, 7:08pm 7

Can you call ldd on the MPI library?

Topic		Replies	Views
CUDA aware MPI works on system but not for Julia Julia at Scale parallel , mpi	30	3024	January 24, 2022
OpenMPI+Threading: Segfault in compilation? General Usage question , multithreading , mpi , segfault	7	929	November 10, 2023
Using views of CuArray with CUDA-aware MPI is extremely slow GPU question , bug , cuda , mpi	14	352	August 5, 2024
Julia crashed - Core dumped General Usage	13	1243	March 13, 2024
CUDA kernel crash very occasionally when MPI.jl is just loaded General Usage question	2	73	July 1, 2024

Error/segfault in basic test of CUDA-aware MPI

Related topics