This is a pre-production deployment of Warehouse. Changes made here affect the production instance of PyPI (
Help us improve Python packaging - Donate today!

GPU-accelerated RIME implementations.

Project Description
# Montblanc

A PyCUDA implementation of the Radio Interferometry Measurement Equation, and a foothill of Mount Exaflop.

## License

Montblanc is licensed under the GNU GPL v2.0 License.

## Requirements

- PyCUDA 2015.1.3
- A Kepler NVIDIA GPU for more recent functionality.

## Installation

Certain pre-requisites must be installed, prior to calling the `` script.

### Pre-requisites

- **[libffi][libffi]** development files, required by cffi. On ubuntu 14.04, you can run:

$ sudo apt-get install libffi-dev

- **[python-casacore][pyrap]** which depends on **[casacore2][casacore]**.
On Ubuntu 14.04 add the [radio astronomy PPA][radio-astro-ppa] and get the binaries from there:

$ sudo apt-get install software-properties-common
$ sudo add-apt-repository ppa:radio-astro/main
$ sudo apt-get update
$ sudo apt-get install python-casacore

### Installing the package

It is strongly recommended that you perform the install within a [Virtual Environment](
If not, consider adding the `--user` flag to the following
pip and python commands to install within your home directory.

First install [numpy] due to this [issue](

$ pip install numpy

Then, run

$ python build

to build the package and

$ python install

to install the package.

### Possible Issues

- **[PyCUDA 2015.1.3][pycuda]**. will attempt to install this automatically,
but this might not work if you have a non-standard CUDA install location. It's worth running

$ python -c 'import pycuda.autoinit'

to check if your pycuda can talk to the NVIDIA driver. If not, manually download
and install [PyCUDA][pycuda].
- **[cub 1.5.1][cub]**. will attempt to download this from github
and install to the correct directory during install. If this fails do the following:

$ wget -c
$ mv 1.5.1
$ python install

- **[numexpr]**. When running `python install`, if you see an error message like this:

Traceback (most recent call last):
File "", line 154, in <module>
File "", line 146, in setup_package
from numpy.distutils.core import setup
ImportError: No module named numpy.distutils.core

Go back to [Installing the package](#installing-the-package) and install [numpy][numpy].

## Running Tests

Once the libraries have been compiled you should be able to run the

$ cd tests
$ python -c 'import montblanc; montblanc.test()'
$ python -m unittest test_biro_v4.TestBiroV4.test_sum_coherencies_double

which will run the entire test suite or only the specified test case, respectively.
The reported times are for the entire test case with numpy code, and not just the CUDA kernels.

If you're running on an ubuntu laptop with optimus technology, you may have to install bumblebee and run

$ optirun python -c 'import montblanc; montblanc.test()'

## Playing with a Measurement Set

You could also try run

$ cd examples
$ python /home/user/data/WSRT.MS -np 10 -ng 10 -c 100

which sets up things based on the supplied Measurement Set, with 10 point and 10 gaussian sources. It performs 100 iterations of the pipeline.

## Citing Montblanc

If you use Montblanc and find it useful, please consider
citing the related [paper][montblanc-paper-sciencedirect].
A [arXiv][montblanc-paper-arxiv] preprint is available.

The BIRO paper is available at [MNRAS][biro-paper-mnras],
and a [arXiv][biro-paper-arxiv] is also available.

## Caveats

Montblanc is an experimental package, undergoing rapid development.
The plan for 2015 is to iterate on new versions of the BIRO pipeline.

In general, I will avoid making changes to BIRO v2 and v3,
but everything beyond that may be changed, including the basic API residing in
In practice, this means that the interfaces in the base montblanc package will remain stable.
For example:

import montblanc

Everything should be considered unstable and subject to change.
I will make an effort to maintain the, to record any breaking API changes.

Release History

Release History

This version
History Node


Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
montblanc-0.4.0a3.tar.gz (90.4 kB) Copy SHA256 Checksum SHA256 Source Mar 11, 2016

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting