kubeflow/trainer

Enhanced MPI Support in TrainJob

Fechada

#2.751 aberto em 24 de jul. de 2025

 (26 comentários) (3 reações) (0 responsável)Go (964 forks)github user discovery
area/controllerarea/mpiarea/runtimesgood first issuehelp wantedkind/feature

Métricas do repositório

Stars
 (2.112 estrelas)
Métricas de merge de PR
 (Mesclagem média 10d 9h) (75 fundiu PRs em 30d)

Description

What you would like to be added?

We plan to enhance MPI support in TrainJob in these areas:

  • SSH-based node communication
  • Add resource per node support
  • Address the failing launcher issue in the initialization phase (launcher fails until the MPI worker is ready)

Why is this needed?

As we discussed in the WG call yesterday, this will be included in Kubeflow Trainer ROADMAP 2025.

/cc @kubeflow/wg-training-leads @astefanutti @franciscojavierarceo @kramaranya @rudeigerc @mahdikhashan @Doris-xm @Harshal292004 @saileshd1402 @kannon92 @deepanker13

Love this feature?

Give it a 👍 We prioritize the features with most 👍

Guia do colaborador