SLATE - Software for Linear Algebra Targeting Exascale

SLATE is a modern C++ library providing dense linear algebra capabilities for high-performance computing systems with multi-core processors and hardware accelerators. SLATE provides coverage of existing ScaLAPACK functionality and is designed as a replacement for ScaLAPACK, which cannot be adequately retrofitted for modern accelerated architectures.

🚀 Quick Start

Get started with SLATE in minutes. Learn installation, basic usage, and run your first example.

Getting Started

📖 User Guide

Comprehensive guide covering matrices, operations, and all SLATE features.

Overview

📚 API Reference

Complete API documentation for all SLATE functions and classes.

slate

💡 Examples

Standalone examples demonstrating SLATE features and usage.

Overview

🧪 Testing & Tuning

Testing suite, benchmarking, and performance tuning guidance.

Testing and Tuning

Key Features

Distributed Computing: Efficient 2D block-cyclic distribution using MPI
GPU Acceleration: Support for NVIDIA (CUDA), AMD (ROCm), and Intel (SYCL) GPUs
Modern C++ API: Clean, type-safe interface with simplified naming
ScaLAPACK Compatible: Easy transition with compatible data layouts and API
Comprehensive Coverage: Linear systems, least squares, eigenvalues, SVD, and more
Mixed Precision: Iterative refinement with lower precision for performance

Quick Example

#include <slate/slate.hh>

// Create matrices on 2x2 process grid
slate::Matrix<double> A(n, n, nb, p, q, MPI_COMM_WORLD);
slate::Matrix<double> B(n, nrhs, nb, p, q, MPI_COMM_WORLD);

// Allocate and initialize tiles
A.insertLocalTiles();
B.insertLocalTiles();

// Solve AX = B using LU factorization
slate::lu_solve(A, B);

Resources

GitHub: https://github.com/icl-utk-edu/slate/
Website: https://icl.utk.edu/slate/
User Forum: https://groups.google.com/a/icl.utk.edu/g/slate-user

Indices and Tables

Citation

@techreport{gates2020slate,
    author={Gates, Mark and Charara, Ali and Kurzak, Jakub and YarKhan, Asim
            and Al Farhan, Mohammed and Sukkari, Dalal and Burgess, Treece
            and Lindquist, Neil and Dongarra, Jack},
    title={{SLATE} Users' Guide, {SWAN} No. 10},
    institution={Innovative Computing Laboratory, University of Tennessee},
    year={2020},
    month={July},
    number={ICL-UT-19-01},
    note={revision 2023-11},
    url={https://www.icl.utk.edu/publications/swan-010},
}