AnySparse

A streamlined, directly compilable port of clSPARSE with a focus on simplicity and broad OpenCL device support. This project, maintained by Prof. Jinchuan Tang (jctang@gzu.edu.cn), is a derivative of the original clSPARSE library. Its primary goal is to eliminate the complexities of the original build system, offering a straightforward "clone and compile" experience. The library supports any OpenCL 1.2+ capable device and is designed for users who need sparse linear algebra operations without the overhead of managing complex build dependencies.

Features

Sparse BLAS Level 1: Operations like axpy, dot, norms.
Sparse BLAS Level 2: Sparse Matrix-Vector multiplication (SpMV).
Sparse BLAS Level 3: Sparse Matrix-Matrix multiplication (SpMM).
Formats: Conversion between CSR, COO, ELL, and other sparse storage formats.
Solvers: Triangular solvers (SpTrSV, SpTrSM).
Preconditioners: Basic preconditioners like diagonal scaling.
Robust support for single-precision
Planned SpMM double-precision
Iterative CG and BiCGStab

Project Activity

See All Activity >

Follow AnySparse

AnySparse Web Site

Other Useful Business Software

Build Securely on AWS with Proven Frameworks

Lay a foundation for success with Tested Reference Architectures developed by Fortinet’s experts. Learn more in this white paper.

Moving to the cloud brings new challenges. How can you manage a larger attack surface while ensuring great network performance? Turn to Fortinet’s Tested Reference Architectures, blueprints for designing and securing cloud environments built by cybersecurity experts. Learn more and explore use cases in this white paper.

Download Now

Rate This Project