KBLAS: an optimized library for dense matrix-vector multiplication on GPU accelerators

DOI10.1145/2818311MaRDI QIDQ5270751zbMATH OpenOpenAlexWikidataFDO

Authors Ahmad Abdelfattah, Hatem Ltaief, D. E. Keyes

Publication date 30 June 2017

Published in ACM Transactions on Mathematical Software (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1410.1726

zbMATH Keywords

basic linear algebra subroutinesCUDA optimizationsGPU acceleratorsmemory-bound kernels

Mathematics Subject Classification ID

Numerical linear algebra (65Fxx) Numerical algorithms for specific classes of architectures (65Y10) Packaged methods for numerical algorithms (65Y15) Complexity and performance of numerical algorithms (65Y20)

Abstract: KBLAS is a new open source high performance library that provides optimized kernels for a subset of Level 2 BLAS functionalities on CUDA-enabled GPUs. Since performance of dense matrix-vector multiplication is hindered by the overhead of memory accesses, a double-buffering optimization technique is employed to overlap data motion with computation. After identifying a proper set of tuning parameters, KBLAS is able to efficiently run on various GPU architectures across different generations, avoiding the time-consuming step of code rewriting, while still being compliant with the standard BLAS API. Another advanced optimization technique allows to ensure coalesced memory access when dealing with submatrices, especially in the context of high level dense linear algebra algorithms. All four precisions KBLAS kernels have been leveraged to multi-GPUs environment, which requires the introduction of new APIs to ease users' experiences on these challenging systems. The KBLAS performance outperforms existing state-of-the-art implementations on all matrix sizes, achieves asymptotically up to 50% and 60% speedup on single GPU and multi-GPUs systems, respectively, and validates our performance model. A subset of KBLAS high performance kernels has been integrated into NVIDIA's standard BLAS implementation (cuBLAS) for larger dissemination, starting version 6.0.

Recommendations

• Accelerating GPU kernels for dense linear algebra
• Redesigning triangular dense matrix computations on GPUs
• Optimizing sparse matrix-matrix multiplication for the GPU
• High-performance matrix-matrix multiplications of very small matrices
• Batched triangular dense linear algebra kernels for very small matrix sizes on GPUs

Cites work

• Accelerating GPU kernels for dense linear algebra
• Accelerating the reduction to upper Hessenberg, tridiagonal, and bidiagonal forms through hybrid GPU-based computing
• Basic Linear Algebra Subprograms for Fortran Usage
• LAPACK Users' Guide

Cited in

(5)

• Accelerated dimension-independent adaptive metropolis
• Accelerating GPU kernels for dense linear algebra
• Hierarchical algorithms on hierarchical architectures
• The BLAS API of BLASFEO: optimizing performance for small matrices
• A hybrid CPU-GPU paradigm to accelerate reactive CFD simulations

Describes a project that uses

Uses Software

• LAPACK
• BLAS
• CUDA
• CUBLAS
• OpenCL
• STREAM
• MAGMA
• PLASMA
• CULA
• STREAM benchmark
• KBLAS

This page was built for publication: KBLAS: an optimized library for dense matrix-vector multiplication on GPU accelerators

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5270751)