SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation

Sparse matrix-sparse matrix multiplication (SpMSpM) is crucial in many fields such as scientific computing, sparse linear algebra, and machine learning due to its computational complexity in the large and extremely sparse datasets. Various applications dealing with the sparse matrix show a variety o...

Full description

Saved in:
Bibliographic Details
Main Authors: Gwanghwi Seo, Sungju Ryu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10921684/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849770687743393792
author Gwanghwi Seo
Sungju Ryu
author_facet Gwanghwi Seo
Sungju Ryu
author_sort Gwanghwi Seo
collection DOAJ
description Sparse matrix-sparse matrix multiplication (SpMSpM) is crucial in many fields such as scientific computing, sparse linear algebra, and machine learning due to its computational complexity in the large and extremely sparse datasets. Various applications dealing with the sparse matrix show a variety of sparse matrix patterns, so the inner product, outer product, and Gustavson (row-wise) methods have been selectively used for the acceleration of the sparse matrix computation. Previous works determine a fixed dataflow before the computation. However, such an approach cannot optimize all the input matrice types having various data patterns. To address these limitations, we propose a SpecBoost, a method that dynamically selects an optimal tile-level SpMSpM dataflow by analyzing the sparsity pattern within each matrix tile and speculating the best tiled dataflow scheme before the computational stage. We compared our method with the widely known previous methods (CSSpa, ExTensor, MatRaptor), and experimental results show that on average our method reduced memory accesses by a factor of (<inline-formula> <tex-math notation="LaTeX">$4.01\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$2.86\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$2.22\times $ </tex-math></inline-formula>) and boosts the performance of prior works over the baseline by (<inline-formula> <tex-math notation="LaTeX">$4.62\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$2.40\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$1.59\times $ </tex-math></inline-formula>).
format Article
id doaj-art-44ff33efd7fa46ffb3b0696a5b49e4d8
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-44ff33efd7fa46ffb3b0696a5b49e4d82025-08-20T03:02:55ZengIEEEIEEE Access2169-35362025-01-0113455684557610.1109/ACCESS.2025.355041410921684SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow SpeculationGwanghwi Seo0https://orcid.org/0009-0001-9255-3184Sungju Ryu1https://orcid.org/0000-0002-0254-391XDepartment of Electronic Engineering, Sogang University, Seoul, Republic of KoreaDepartment of System Semiconductor Engineering, Sogang University, Seoul, Republic of KoreaSparse matrix-sparse matrix multiplication (SpMSpM) is crucial in many fields such as scientific computing, sparse linear algebra, and machine learning due to its computational complexity in the large and extremely sparse datasets. Various applications dealing with the sparse matrix show a variety of sparse matrix patterns, so the inner product, outer product, and Gustavson (row-wise) methods have been selectively used for the acceleration of the sparse matrix computation. Previous works determine a fixed dataflow before the computation. However, such an approach cannot optimize all the input matrice types having various data patterns. To address these limitations, we propose a SpecBoost, a method that dynamically selects an optimal tile-level SpMSpM dataflow by analyzing the sparsity pattern within each matrix tile and speculating the best tiled dataflow scheme before the computational stage. We compared our method with the widely known previous methods (CSSpa, ExTensor, MatRaptor), and experimental results show that on average our method reduced memory accesses by a factor of (<inline-formula> <tex-math notation="LaTeX">$4.01\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$2.86\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$2.22\times $ </tex-math></inline-formula>) and boosts the performance of prior works over the baseline by (<inline-formula> <tex-math notation="LaTeX">$4.62\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$2.40\times $ </tex-math></inline-formula>, <inline-formula> <tex-math notation="LaTeX">$1.59\times $ </tex-math></inline-formula>).https://ieeexplore.ieee.org/document/10921684/Matrix tilingsparse matrix multiplicationtile-level dataflow speculatormatrix sampling with threshold
spellingShingle Gwanghwi Seo
Sungju Ryu
SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation
IEEE Access
Matrix tiling
sparse matrix multiplication
tile-level dataflow speculator
matrix sampling with threshold
title SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation
title_full SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation
title_fullStr SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation
title_full_unstemmed SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation
title_short SpecBoost: Accelerating Tiled Sparse Matrix Multiplication via Dataflow Speculation
title_sort specboost accelerating tiled sparse matrix multiplication via dataflow speculation
topic Matrix tiling
sparse matrix multiplication
tile-level dataflow speculator
matrix sampling with threshold
url https://ieeexplore.ieee.org/document/10921684/
work_keys_str_mv AT gwanghwiseo specboostacceleratingtiledsparsematrixmultiplicationviadataflowspeculation
AT sungjuryu specboostacceleratingtiledsparsematrixmultiplicationviadataflowspeculation