Implementing Sparse Matrix-Vector Multiplication on Throughput-Oriented Processors
Nathan Bell, Michael Garland · SC 2009
Open reference学术参考
本页把论文、项目和延伸阅读分开整理，方便读者快速建立“这个项目参考了谁、站在什么技术谱系上”的理解。
核心论文
Merge-based Parallel Sparse Matrix-Vector Multiplication
Duane Merrill, Michael Garland · SC 2016
Open referenceAutomatic Tuning of the Sparse Matrix Vector Product on GPUs Based on the ELL-R-T Format
Fernando Vázquez et al. · Concurrency and Computation 2011
Open reference代表性项目
NVIDIA cuSPARSE
Vendor baseline and API reference for sparse GPU primitives
Open referenceGinkgo
Production-quality sparse linear algebra library with strong docs
Open referenceModernGPU
Useful for understanding scan / merge / partitioning techniques on GPU
Open referenceSuiteSparse Matrix Collection
Representative real-world sparse matrices for benchmark reasoning
Open reference如何使用这些参考
- 先读 Bell & Garland，理解 GPU SpMV 的经典问题定义。
- 再看 Merrill & Garland，理解 Merge Path 在不规则负载中的价值。
- 对照 cuSPARSE / Ginkgo / SuiteSparse，把本项目放回真实工程生态里看。