sasmodels Installation

Sasmodels can be installed using a simple pip installation:

# Won't work yet! For now follow the developer instructions below.
$ pip install sasmodels

There are a number of dependent python packages which need to be installed separately:

numpy
scipy
opencl (for speed if your system has OpenCL drivers)
tinycc (windows only, if no C compiler and no OpenCL)

When using sasmodels to fit data:

sasview (for loading data)
bumps (for fitting)
matplotlib (for displaying results)

For development and testing it is handy to have some additional packages:

nose (for testing)
xmlrunner (for testing)
matplotlib (for showing models)
docutils (for showing model docs)
sphinx (for building docs)
wx ([optional] for adjusting parameters interactively)

Developer Installation

Developers will need to clone the sasmodels reposistory from github:

$ git clone https://github.com/sasview/sasmodels.git

or if you have a github account:

$ git clone git@github.com:sasview/sasmodels.git

Sasmodels can be run in-place by including the sasmodels directory on the python path. There are numerous ways to do this which will not be outlined here. It will be easiest to install the package in "develop" mode using:

$ cd sasmodels
$ python setup.py develop

This will allow you to edit the files in the package and have the changes show up immediately in python the next time you load your program.

OpenCL Installation

Warning! GPU devices do not in general offer the same level of memory protection as CPU devices. If your code attempts to write outside allocated memory buffers unpredicatable behaviour may result (eg, your video display may freeze, or your system may crash, etc). Do not install OpenCL drivers without first checking for known issues (eg, some computer manufacturers install modified graphics drivers so replacing these may not be a good idea!). If in doubt, seek advice from an IT professional before proceeding further.

GPU Selection

sasmodels evaluations can run on your graphics card (GPU) or they can run on the processor (CPU). In general, calculations performed on the GPU will run faster.

To run on the GPU, your computer must have OpenCL drivers installed. For information about OpenCL installation see this :ref:`opencl-installation` guidance.

Where the model is evaluated is a little bit complicated. If the model has the line single=False then it requires double precision. If the GPU is single precision only, then it will try running via OpenCL on the CPU. If the OpenCL driver is not available for the CPU then it will run as a normal program on the CPU.

For models with a large number of parameters or with a lot of code, the GPU may be too small to run the program effectively. In this case, you should try simplifying the model, maybe breaking it into several different modules so that you don't need IF statements in your code. If it is still too big, you can set opencl=False in the model file and the model will only run as a normal program on the CPU. This will not usually be necessary.

import pyopencl as cl
cl.create_some_context()