A V library for AI and high-performance scientific computing with pure-V BLAS/LAPACK implementations.
VSL (The V Scientific Library) is a comprehensive scientific computing library for the V programming language that provides tools for artificial intelligence, high-performance numerical computations, and data visualization. It solves the problem of performing complex scientific calculations in V by offering pure-V implementations of essential algorithms alongside optional integrations with optimized external libraries.
V developers working on AI, machine learning, scientific research, or numerical computing projects who need high-performance mathematical operations and data analysis capabilities.
Developers choose VSL for its unique combination of dependency-free pure-V implementations, optional high-performance backends, and comprehensive scientific modules—all within the V ecosystem's simplicity and safety guarantees.
V library to develop Artificial Intelligence and High-Performance Scientific Computations
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Offers pure-V implementations of BLAS/LAPACK with zero external dependencies, enabling high-performance linear algebra without installing system libraries like OpenBLAS.
Provides optional integration with OpenBLAS, LAPACK, MPI, and OpenCL via compilation flags, allowing users to trade simplicity for maximum performance when needed.
Includes a wide range of tools from linear algebra and machine learning to data visualization and statistical analysis, all within a single library ecosystem.
Works across all platforms supported by V, ensuring portability for scientific applications without backend compatibility issues.
Features comprehensive test coverage and benchmark suites, as highlighted in CI badges, indicating a focus on reliable, tested implementations.
The pure-V QR decomposition (geqrf/orgqr) is still being aligned and temporarily skipped, forcing reliance on C backends for correctness, as admitted in the IMPORTANT warning.
Being tied to the V language, it lacks the extensive community, documentation, and third-party integrations of established alternatives like NumPy or Julia's SciML.
Using optimized backends requires installing system libraries (e.g., OpenBLAS, OpenCL) and managing compilation flags, which adds overhead compared to drop-in solutions.
Pure-V implementations, while competitive, may not consistently match the peak performance of decades-optimized C/Fortran libraries, as benchmark comparisons imply potential gaps.