Installation guide
Myokit runs on Python 3.6 or newer, is pip-installable and requires the CVODES library to be installed. For more detailed instructions, see below:
If you're planning to use OpenCL for multi-cellular simulations, you might want to consult the OpenCL installation hints given below.
Solving installation issues
There are a few Myokit tricks that can help debug common installation problems. To get an overview of the dependencies Myokit detects on the system, use
$ python -m myokit systemTo see which compiler Myokit detects, use
$ python -m myokit compilerTo see which Sundials version can be found, use
$ python -m myokit sundialsFinally, to get information about the OpenCL devices and drivers detected, use
$ python -m myokit opencl
Configuration files
Myokit stores user-specific settings in configuration files
using the extension .ini.
These files are located in ~/.config/myokit,
where ~ denotes the user home directory.
(When using WinPython, this might be
WinPython\settings\.config\myokit).
Single-cell simulations
Single-cell simulations in Myokit are run using the CVODES
integrator from Sundials. On most linux systems, this can
be installed through a package manager. If you have
installed Sundials, but Myokit has trouble finding its
header files (with filenames ending in .h), please check
the paths specified in myokit.ini (see
above
).
If you get an error saying the shared libraries can't be
found, please check your LD settings on Linux
and Mac or your Path environment variable on
Windows.
For more help, contact Michael Clerx (michael@myokit.org) or use the issues page.
Multi-cell simulations with OpenCL
The installation procedure for OpenCL is specific to the used hardware and operating system. In general, you'll need to install two things:
- An OpenCL implementation. Different implementations are available from different vendors, usually bundled as an SDK.
- An OpenCL supporting driver for the hardware you want to run the code on (Modern drivers tend to support OpenCL already, so you might be able to skip this step).
OpenCL drivers are specific to the device (CPU or GPU) you're using. It is possible to install drivers for several devices side by side. OpenCL implementations are device independent, and you can use the version supplied by any vendor.
For details on installing OpenCL, check with the vendor of your hardware.
Once OpenCL has been installed, Myokit will need to know
where it can find the required binaries and header files.
The standard configuration files makes a few common
guesses, but if Myokit says it cannot find
CL/cl.h, you'll need to check the paths given
in myokit.ini (see
above
).
To see if Myokit can detect the OpenCL drivers and devices, use
$ python -m myokit openclIf that works, you can select the "platform" (OpenCL software implementation) and device to execute on using
$ python -m myokit opencl-select
If you get errors saying the shared libraries cannot be
found, please check your LD settings on Linux
and Mac or the environment variable Path on
Windows.
For more help, contact Michael Clerx (michael@myokit.org) or checkout out the issues page.