Compiling HOOMD-blue in RHEL/CentOS

Note: This documentation was written and tested in RHEL/CentOS 5.5. It may not work for other versions.

This page assumes that you have a standard terminal window open. Commands to run will be indicated as below:

 $ echo hello
hello

" $ " indicates a shell prompt. As demonstrated above, if you type "echo hello", then you should see the same output obtained above on the next line: "hello"

You must be the root user to install software. Become root by executing the following command.

 $ su -

Development environment

Install g++, boost, python headers, zlib headers, and git.

 $ yum install gcc-c++ boost boost-devel python-devel zlib-devel git-all

CMake is not available via yum, so download cmake-?.?.?-Linux-i386.sh from http://www.cmake.org/ . Replace ?.?.? with the most current version. The following terminal command is an example that downloads version 2.8.3

 $ wget http://www.cmake.org/files/v2.8/cmake-2.8.3-Linux-i386.sh

Install CMake by executing

 $ bash cmake*.sh --prefix=/usr/local --exclude-subdir

CUDA environment

Download the current version of the NVIDA CUDA toolkit from http://developer.nvidia.com/object/gpucomputing.html . The following terminal command is an example that downloads version 3.2 for RHEL/CentOS 5.5 x86_64

 $ wget http://developer.download.nvidia.com/compute/cuda/3_2_prod/toolkit/cudatoolkit_3.2.16_linux_64_rhel5.5.run

Install the CUDA toolkit by executing

 $ bash cudatoolkit*.run

and following the onscreen prompts. Accept the default installation directory.

Add CUDA into the executable and library search paths by creating the following files. Create /etc/ld.so.conf.d/cuda.conf and add the following lines

/usr/local/cuda/lib
/usr/local/cuda/lib64

Then run

 $ ldconfig

Create /etc/profile.d/cuda.sh and add the following lines

if ! echo ${PATH} | /bin/grep -q /usr/local/cuda/bin ; then
    PATH=/usr/local/cuda/bin:${PATH}
fi

You will need to logout and back in for the path setting to take effect.

Download the latest drivers for your GPU from http://www.nvidia.com The following terminal command is an example that downloads version 260.19.21 for linux x86_64. If you do not have an NVIDIA GPU, proceed to the section Compile hoomd.

 $ wget http://us.download.nvidia.com/XFree86/Linux-x86_64/260.19.21/NVIDIA-Linux-x86_64-260.19.21.run

To install the driver, you must exit first logout of your X-windows session. At the login screen, press ctrl-alt-f1 to reach a text console. Login as root and execute the following commands

 $ init 3
 $ yum install kernel-devel-`uname -r`
 $ bash NVIDIA-Linux-*.run

and follow the onscreen prompts to install the driver. Reboot the computer to load the new driver and re-enter the GUI login screen.

 $ shutdown now -r

Compile hoomd

Log out of any root sessions that have been opened to install software above. Software should be compiled in a standard user account.

See Compile hoomd for instructions on compiling hoomd.