Autotuning GEMM Kernels for the Fermi GPU
Authors:
- 3 January 2012
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 23 (11) , 2045-2057
Abstract
In recent years, the use of graphics chips has been recognized as a viable way of accelerating scientific and engineering applications, even more so since the introduction of the Fermi architecture by NVIDIA, with features essential to numerical computing, such as fast double precision arithmetic and memory protected with error correction codes. Being the crucial component of numerical software packages, such as LAPACK and ScaLAPACK, the general dense matrix multiplication routine is one of the more important workloads to be implemented on these devices. This paper presents a methodology for producing matrix multiplication kernels tuned for a specific architecture, through a canonical process of heuristic autotuning, based on generation of multiple code variants and selecting the fastest ones through benchmarking. The key contribution of this work is in the method for generating the search space; specifically, pruning it to a manageable size. Performance numbers match or exceed other available implementations.Keywords
This publication has 29 references indexed in Scilit:
- A fast GEMM implementation on the cypress GPUACM SIGMETRICS Performance Evaluation Review, 2011
- An Improved Magma Gemm For Fermi Graphics Processing UnitsThe International Journal of High Performance Computing Applications, 2010
- Complex version of high performance computing LINPACK benchmark (HPL)Concurrency and Computation: Practice and Experience, 2009
- Optimizing matrix multiplication for a short-vector SIMD architecture – CELL processorParallel Computing, 2009
- Anatomy of high-performance matrix multiplicationACM Transactions on Mathematical Software, 2008
- The LINPACK Benchmark: past, present and futureConcurrency and Computation: Practice and Experience, 2003
- Automated empirical optimizations of software and the ATLAS projectParallel Computing, 2001
- GEMM-based level 3 BLASACM Transactions on Mathematical Software, 1998
- Single event upset at ground levelIEEE Transactions on Nuclear Science, 1996
- Efficient complex matrix multiplicationIEEE Transactions on Computers, 1988