Questions tagged [ucx]

UCX claims to be "an open-source production grade communication framework for data centric and high-performance applications". Use this tag for questions about using the framework, but be advised that there may be only a handful of experienced users in the SO community.

UCX claims to be "an open-source production grade communication framework for data centric and high-performance applications". Its source is maintained on github, and some documentation is available on the OpenUCX site.

7 questions
2
votes
0 answers

Using UCX protocol Dask Distributed

I would like to take advantage of the InfiniBand network to connect Dask Client and the workers and scheduler (especially between the clients and workers -not necessary with GPUs- as I scatter some data directly to workers). I am using the CLI to…
Mitchou
  • 37
  • 5
2
votes
1 answer

UCX warn unexpected tag-receive

What can the following be due to / how to debug it? it happens when closing my MPI application [1612979755.727913] [compute-0-9:21112:0] tag_match.c:61 UCX WARN unexpected tag-receive descriptor 0x2b2bf64cdbc0 was not matched
ATK
  • 1,296
  • 10
  • 26
1
vote
1 answer

How to enable CUDA Aware OpenMPI?

I'm using OpenMPI and I need to enable CUDA aware MPI. Together with MPI I'm using OpenACC with the hpc_sdk software. Following https://www.open-mpi.org/faq/?category=buildcuda I downloaded and installed UCX (not gdrcopy, I haven't managed to…
Steve
  • 89
  • 1
  • 6
1
vote
3 answers

Please use compiler that supports __attribute__((constructor))

I just compiled my own version of gcc/9.2.0 using gcc/4.8.2. After successful compilation and installation of gcc/9.2.0 I try compiling ucx-1.5.1. When I try to run the ucx configure script I get the following message "checking…
L.H
  • 33
  • 2
0
votes
0 answers

OSU benchmark hang after finish printed the result when using UCX in the container

I am experimenting with using UCX to provide more portable MPI app containers with performance. So I want to compare between using UCX replacement method which mount system built UCX into container at runtime(sure with other interconnect libraries)…
tk11
  • 1
0
votes
1 answer

mm_xpmem.c UCX error failed to attach xpmem

I am running analysis on a cluster and internally I am spawning some processes. Most of the times it works, but sometimes I get following error: mm_xpmem.c:135 UCX ERROR failed to attach xpmem apid 0x600005c0e offset 0x2b8cb9183000 length 12288:…
Pavan
  • 133
  • 7
0
votes
1 answer

How to install ucp module in python? [dask]

I am trying to run on a dask cluster using the ucx protocol. I got to know from my admins that the protocol is installed as expected. I however receive the following error when I am trying to switch to the ucx protocol: from ucp import get_config,…
NiRvanA
  • 105
  • 1
  • 1
  • 8