pytorch/xla

xs.make_mesh

Open

#8,838 opened on Mar 14, 2025

View on GitHub
 (1 comment) (0 reactions) (1 assignee)C++ (575 forks)github user discovery
enhancementgood first issue

Repository metrics

Stars
 (2,786 stars)
PR merge metrics
 (PR metrics pending)

Description

🚀 Feature

We should provide a xs.make_mesh method similar to JAX: https://docs.jax.dev/en/latest/_autosummary/jax.make_mesh.html

Motivation

The ordering of devices affects the performance of collective operations:

  • Weights are sharded differently and the layout of tensors after an all-gather is different.
  • The performance of a collective op itself depends on what network connections it uses.

Pitch

The xs.make_mesh method should attempt to create a reasonably optimized device ordering given a mesh shape referencing how JAX does it.

We could also deviate from JAX and add additional options that fit our own needs.

cc @yaochengji

Contributor guide

Research direction
Analyze JAX's make mesh implementation and understand how PyTorch/XLA currently handles device mesh and collective operations. Identify the key performance factors and propose an optimized device ordering strategy.
Tech stack
python
Domain
backendperformance
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
PythonC++PyTorch
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40