ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2404.15559
56
0
v1v2 (latest)

Low-Bandwidth Matrix Multiplication: Faster Algorithms and More General Forms of Sparsity

23 April 2024
Chetan Gupta
Janne H. Korhonen
Jan Studený
Jukka Suomela
Hossein Vahidi
ArXiv (abs)PDFHTML
Abstract

In prior work, Gupta et al. (SPAA 2022) presented a distributed algorithm for multiplying sparse n×nn \times nn×n matrices, using nnn computers. They assumed that the input matrices are uniformly sparse--there are at most ddd non-zeros in each row and column--and the task is to compute a uniformly sparse part of the product matrix. The sparsity structure is globally known in advance (this is the supported setting). As input, each computer receives one row of each input matrix, and each computer needs to output one row of the product matrix. In each communication round each computer can send and receive one O(log⁡n)O(\log n)O(logn)-bit message. Their algorithm solves this task in O(d1.907)O(d^{1.907})O(d1.907) rounds, while the trivial bound is O(d2)O(d^2)O(d2). We improve on the prior work in two dimensions: First, we show that we can solve the same task faster, in only O(d1.832)O(d^{1.832})O(d1.832) rounds. Second, we explore what happens when matrices are not uniformly sparse. We consider the following alternative notions of sparsity: row-sparse matrices (at most ddd non-zeros per row), column-sparse matrices, matrices with bounded degeneracy (we can recursively delete a row or column with at most ddd non-zeros), average-sparse matrices (at most dndndn non-zeros in total), and general matrices.

View on arXiv
Comments on this paper