A Generalized Framework for Auto-tuning Stencil Computations
Conference
·
OSTI ID:962935
This work introduces a generalized framework for automatically tuning stencil computations to achieve superior performance on a broad range of multicore architectures. Stencil (nearest-neighbor) based kernels constitute the core of many important scientific applications involving block-structured grids. Auto-tuning systems search over optimization strategies to find the combination of tunable parameters that maximizes computational efficiency for a given algorithmic kernel. Although the auto-tuning strategy has been successfully applied to libraries, generalized stencil kernels are not amenable to packaging as libraries. Studied kernels in this work include both memory-bound kernels as well as a computation-bound bilateral filtering kernel. We introduce a generalized stencil auto-tuning framework that takes a straightforward Fortran expression of a stencil kernel and automatically generates tuned implementations of the kernel in C or Fortran to achieve performance portability across diverse computer architectures.
- Research Organization:
- Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA (US)
- Sponsoring Organization:
- Computational Research Division
- DOE Contract Number:
- AC02-05CH11231
- OSTI ID:
- 962935
- Report Number(s):
- LBNL-2078E
- Country of Publication:
- United States
- Language:
- English
Similar Records
A Generalized Framework for Auto-tuning Stencil Computations
Stencil computation optimization and auto-tuning on state-of-the-art multicore architectures.
Stencil Computation Optimization and Auto-tuning on State-of-the-Art Multicore Architectures
Conference
·
Mon Aug 24 00:00:00 EDT 2009
·
OSTI ID:1407077
Stencil computation optimization and auto-tuning on state-of-the-art multicore architectures.
Conference
·
Thu Nov 20 23:00:00 EST 2008
·
OSTI ID:1407060
Stencil Computation Optimization and Auto-tuning on State-of-the-Art Multicore Architectures
Conference
·
Fri Aug 22 00:00:00 EDT 2008
·
OSTI ID:964371