Skip to content

Latest commit

 

History

History
35 lines (27 loc) · 1.06 KB

README.md

File metadata and controls

35 lines (27 loc) · 1.06 KB

SpComm3D: A Framework for Enabling Sparse Communication in 3D Sparse Kernels

Overview

SpComm3D is an MPI-based framework for building distributed-memeory 3D sparse kernels for CPUs and GPUs. The main feature of SpComm3D is that it efficiently uses sparsity-aware communication in 3D sparse kernels instead of sparsity-agnostic communication.

Currently, SpComm3D contains sparsity-aware implementations for 3D algorithms of two important ML sparse kernels that occur frequently in Graph/Hypergraph Neural Networks:

  • Sampled Dense-Dense Matrix Multiplication (SDDMM)
  • Sparse Matrix-Dense Matrix Multiplication (SpMM)

To build conv2bin tool:

make conv2bin

To make benchSddmm:

make benchSddmm  

Authors

Nabil Abubaker, Torsten Hoefler

Citation:

@misc{abubaker2024spcomm3d,
      title={SpComm3D: A Framework for Enabling Sparse Communication in 3D Sparse Kernels}, 
      author={Nabil Abubaker and Torsten Hoefler},
      year={2024},
      eprint={2404.19638},
      archivePrefix={arXiv},
      primaryClass={cs.DC}
}