dmlc/dgl

The ALG2 for cusparsespmm under performed when feature size is small

Open

#4,090 opened on Jun 7, 2022

View on GitHub
 (2 comments) (0 reactions) (0 assignees)Python (12,665 stars) (2,928 forks)batch import
help wanted

Description

🐛 Bug

@yzh119 report the ALG2 seems performed badly for cusparsespmm when the feature size is small.

To Reproduce

Steps to reproduce the behavior:

Expected behavior

Environment

  • DGL Version (e.g., 1.0):
  • Backend Library & Version (e.g., PyTorch 0.4.1, MXNet/Gluon 1.3):
  • OS (e.g., Linux):
  • How you installed DGL (conda, pip, source):
  • Build command you used (if compiling from source):
  • Python version:
  • CUDA/cuDNN version (if applicable):
  • GPU models and configuration (e.g. V100):
  • Any other relevant information:

Additional context

Contributor guide

Tech stack
pythoncpppytorch
Domain
backendperformance
Issue type
performance
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.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
unclear
Prerequisites
CUDA programmingsparse matrix operationsDGL internals
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20
Research direction
The issue describes a performance regression in the cusparsespmm algorithm ALG2 for small feature sizes. To investigate, examine the CUDA kernel implementation in the DGL source code, likely in files related to sparse matrix operations. The reporter is @yzh119, and there are no linked PRs or further discussion, so the first step would be to reproduce the issue using a simple benchmark and compare ALG2 with other algorithms.