triton-lang/triton

Implement BatchNorm in triton

Open

#900 opened on Nov 22, 2022

View on GitHubContributor guide
 (6 comments) (0 reactions) (0 assignees)MLIR (3,126 forks)github user discovery
help wanted

Repository metrics

Stars
 (19,982 stars)
PR merge metrics
 (Avg merge 2d 18h) (185 merged PRs in 30d)

Description

batchnorm(bn) is very popular in CV, almost every conv op will be followed by bn. I see layernorm in triton achieved best HBM bandwidth. So I'm curious about implement batchnorm in triton.

My questions:

  1. is there any chance that triton implemented bn faster than pytorch?
  2. may you guys give more details on layernorm's triton implementation? about why it achieved so amazing bandwidth

Contributor guide

Research direction
Implement a batchnorm kernel in Triton, benchmark against PyTorch's batchnorm, and analyze bandwidth utilization following the layernorm tutorial.
Tech stack
python
Domain
backendmachine learning
Issue type
Research
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.
Fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
PythonTritonCUDA
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40