https://github.com/ucb-bar/lbnl-torch
LBNL TORCH Reference Kernels
https://github.com/ucb-bar/lbnl-torch
Last synced: 3 months ago
JSON representation
LBNL TORCH Reference Kernels
- Host: GitHub
- URL: https://github.com/ucb-bar/lbnl-torch
- Owner: ucb-bar
- Created: 2019-08-13T22:26:47.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2019-08-14T01:27:00.000Z (almost 6 years ago)
- Last Synced: 2025-01-16T01:15:25.772Z (5 months ago)
- Language: C
- Size: 189 KB
- Stars: 2
- Watchers: 10
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README
- Changelog: NEWS
Awesome Lists containing this project
README
/*
* Kernel Reference Testbed: version X
* parallelmotifs.org
*
* Last Modified: September 2010
*
* Include License
*
* Authors:
*
*
*/
About the motifs and kernels:
=============================
In the current release (version X), we include X "kernels" that span
eight different "computational motifs".
Kernels are problem definitions, and we provide a reference implementation
(one particular solution) for each kernel. The kernels are organized by the
primary motif they belong to, and so there are eight motif directories.
The reference implementations are either in C or MATLAB (in some cases,
both). All MATLAB routines corresponding to kernels for a motif are under
sub-directories named MATLAB.
Building the C implementations:
===============================
1. Set the C compiler and flags in Makefile.in.
2. Execute "make" in the top directory to build all the kernel
implementations. Alternately, execute a make in any sub-directory
to build the kernel implementations corresponding to that motif.
Execution:
==========
See the README files in each sub-directory for instructions on
problem generation, kernel execution, and verification.
Just executing the kernel without any command-line arguments
prints out usage details.
References:
===========
Included in the doc directory.
Cite
1. Tech Report
2. HotPar'10 paper
Contact:
========