- Integration of CiNii Books functions for fiscal year 2025 has completed
- Trial version of CiNii Research Knowledge Graph Search feature is available on CiNii Labs
- 【Updated on November 26, 2025】Regarding the recording of “Research Data” and “Evidence Data”
- Incorporated Jxiv preprints from JaLC and adding coverage from NDL Search
A Dynamic Parameter Tuning Method for High Performance SpMM
Bibliographic Information
- Published
- 2021
- Resource Type
- journal article
- Rights Information
-
- http://www.springer.com/tdm
- http://www.springer.com/tdm
- DOI
-
- 10.1007/978-3-030-69244-5_28
- Publisher
- Springer International Publishing
Search this article
Description
Sparse matrix-matrix multiplication (SpMM) is a basic kernel that is used by many algorithms. Several researches focus on various optimizations for SpMM parallel execution. However, a division of a task for parallelization is not well considered yet. Generally, a matrix is equally divided into blocks for processes even though the sparsities of input matrices are different. The parameter that divides a task into multiple processes for parallelization is fixed. As a result, load imbalance among the processes occurs. To balance the loads among the processes, this paper proposes a dynamic parameter tuning method by analyzing the sparsities of input matrices. The experimental results show that the proposed method improves the performance of SpMM for examined matrices by up to 39.5% and 12.3% on average.
Journal
-
- Lecture Notes in Computer Science
-
Lecture Notes in Computer Science 318-329, 2021
Springer International Publishing
- Tweet
Details 詳細情報について
-
- CRID
- 1360009142744487680
-
- ISSN
- 16113349
- 03029743
-
- Article Type
- journal article
-
- Data Source
-
- Crossref
- KAKEN
- OpenAIRE
