Matlab Install¶

Requirements¶

A unix-like system (typically Linux or MacOs X) with a C++ compiler compatible with std=c++14 (e.g. gcc>=7 or clang>=8)
Cmake>=3.10
Matlab>=R2012
Optional: Cuda (>=10.0 is recommended)

Packaged version (recommended)¶

Download and unzip the KeOps library at a location of your choice (denoted here by /path/to):

cd /path/to
wget https://github.com/getkeops/keops/archive/master.zip
unzip master.zip

Note that temporary files will be written into the /path/to/keops-master/keopslab/build folder: this directory must have write permissions.

Manually add the directory /path/to/keops-master/keopslab to you Matlab path, as documented below.
Test your installation.

From source using git¶

Clone the KeOps repository at a location of your choice (denoted here by /path/to):

git clone https://github.com/getkeops/keops.git /path/to/keops

Note that temporary files will be written into the /path/to/keops/keopslab/build folder: this directory must have write permissions.

Manually add the directory /path/to/keops/keopslab to you matlab path: see Set the path
Test your installation.

Set the path¶

There are two ways to tell Matlab that KeOpsLab is now available in /path/to/keops/keopslab:

You can add this folder to your Matlab path once and for all: in the Matlab prompt, type

addpath(genpath('/path/to/keops/keopslab'))
savepath

Otherwise, please add the following line to the beginning of your scripts:

addpath(genpath('/path/to/keops/keopslab'))

Test that everything goes fine¶

Set the path and execute the following piece of code in a Matlab terminal:

x = reshape(1:9,3,[]); y = reshape(3:8,3,[]);

my_conv = keops_kernel('Sum_Reduction(SqNorm2(x-y),1)','x=Vi(0,3)','y=Vj(1,3)');
my_conv(x,y)'

It should return:

ans =
    63
    90

Troubleshooting¶

Verbosity¶

For debugging purposes, you can force a “verbose” compilation mode by setting

verbosity=1

in the file /path/to/keops/keopslab/default_options.m.

Old versions of Cuda¶

When using KeOps with Cuda version 8 or earlier, the compilation of complex formulas may take a very long time (several minutes). This typically happens when computing the derivative or second-order derivatives of a non-trivial function, as in the testShooting.m example script.

This delay is mainly due to the intensive use of modern C++11 templating features, for which the old (<=8) versions of the Cuda nvcc compiler were not optimized. Consequently, if you own a GPU with a compute capability >=3.0, we strongly recommend upgrading to Cuda>=9.

Cmake is not found¶

If an error involving cmake appears, it may be due to an incorrect linking of libstdc++. Try the following: exit Matlab, then type in a terminal

export LD_PRELOAD=$(ldd $( which cmake ) | grep libstdc++ | tr ' ' '\n' | grep /)
matlab

This will reload Matlab with, hopefully, a correct linking for cmake.