NVIDIA/SKA-gpu-multiscale-clean

Name: SKA-gpu-multiscale-clean

Owner: NVIDIA Corporation

Description: Multiscale CLEAN for the square kilometer array using GPUs

Created: 2016-08-08 19:51:25.0

Updated: 2017-06-28 22:16:14.0

Pushed: 2017-06-26 23:18:21.0

Homepage: null

Size: 22

Language: C++

GitHub Committers

UserMost Recent Commit# Commits

Other Committers

UserEmailMost Recent Commit# Commits

README

GPUMultiscaleCLEAN

Multi-scale clean for the GPU and/or CPU

This code is written by Mark Harris, based on work from the ASKAP project, and modified for multi-scale by Levi Barnes. The original code can be found here: https://github.com/askap-benchmarks/

This code requires CUB. It can be found here: https://github.com/NVlabs/CUB Specify CUB_HOME in the Makefile to build.

Execution of the tMultiscaleClean and tHogbomClean benchmarks will require the existence of the point spread function (PSF) image and the dirty image (the image to be cleaned) the working directory. These can be downloaded from here:

Build

To build GPU version, specify CUDA=1 on the make line

> make CUDA=1

To run with GPU-only (i.e. without error checking), specify skipgolden on the command line

> ./tMultiScaleClean skipgolden

This work is supported by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number U24TR002306. This work is solely the responsibility of the creators and does not necessarily represent the official views of the National Institutes of Health.