How does it handle numerical errors?

It includes patterns for numerical stability, such as Kahan compensated summation and stable L2 norm computation to minimize floating-point errors.

Is this compatible with modern C++ features?

The skill is specifically designed for C++20, utilizing modern features like std::span, ranges, and updated memory management patterns.

Does this skill support sparse matrices?

Yes, it includes optimized patterns for Compressed Sparse Row (CSR) matrix formats and SpMV operations.

Can I use this for parallel computing?

Absolutely, it provides specific implementations for MPI-IO and parallel file access using HDF5 for high-performance distributed environments.

C++20 Numerical Patterns

Name: C++20 Numerical Patterns
Author: ysyecust

byysyecust

•

데이터 과학 및 ML

Provides standardized C++20 implementations for high-performance numerical computing, matrix operations, and parallel I/O.

소개

This skill equips Claude with advanced domain knowledge for scientific computing and High-Performance Computing (HPC) development using C++20. It includes optimized patterns for dense and sparse matrix operations, iterative solvers like Conjugate Gradient and GMRES, and techniques for numerical stability such as Kahan summation. Additionally, it provides robust templates for parallel file access using MPI-IO and HDF5, enabling the creation of efficient, production-grade data pipelines for large-scale simulations and data analysis tasks.

주요 기능

Numerically stable algorithms like Kahan summation and stable norm computation
Parallel I/O integration with MPI-IO and HDF5 for HPC workloads
Production-grade iterative solvers including CG and GMRES
2 GitHub stars
Optimized dense and sparse (CSR) matrix operation patterns
Streaming and chunked data pipeline architectures for large datasets

사용 사례

Implementing memory-efficient sparse matrix-vector multiplication (SpMV) in C++
Optimizing I/O performance for large-scale distributed datasets on supercomputers
Developing high-performance physics simulations or finite element analysis tools

소개

주요 기능

Numerically stable algorithms like Kahan summation and stable norm computation
Parallel I/O integration with MPI-IO and HDF5 for HPC workloads
Production-grade iterative solvers including CG and GMRES
2 GitHub stars
Optimized dense and sparse (CSR) matrix operation patterns
Streaming and chunked data pipeline architectures for large datasets

사용 사례

Implementing memory-efficient sparse matrix-vector multiplication (SpMV) in C++
Optimizing I/O performance for large-scale distributed datasets on supercomputers
Developing high-performance physics simulations or finite element analysis tools