Taverna is transitioning to the Apache Incubator. Currently there is no Apache-approved distribution.
Taverna Workbench and Command Line are available in either platform-specific or platform-neutral distributions. We generally recommend the platform-specific distributions as they include any required third-party software.
Platform-specific distributions are available for (and tested on):
- Windows 64-bit (.exe installer) (7, 8, 8.1)
- Windows 32-bit (.exe installer) (7)
- Mac OS X 64-bit (.dmg disk image) (10.8, 10.9)
- Linux 64-bit Debian-based (.deb package) (Ubuntu 12.04.4, Bio-Linux 7)
- Linux 64-bit Redhat-based (.rpm package) (Fedora 20)
Platform-neutral distributions are available as:
- Zip file
The platform-neutral distribution has additional system requirements.
If you want to use the Interaction service from Taverna, for some versions of Linux it may be necessary to install the libgnome-2.0 package or equivalent. This library allows Taverna to open pages in your preferred Web browser and connect to the interactive service being called from the workflow.
It is a good idea for your machine to have at least 2GB of RAM. Taverna will run with less memory, but its performance may be affected.
The Taverna Workbench installations for 64-bit operating systems will use by default up to 1 GB of working memory, plus up to 400 MB used by the Java runtime. Installations for 32-bit operating systems and platform-independent installations will use up to 700 MB of working memory, plus up to 200 MB used by Java runtime.
Reduce memory usage¶
If Taverna runs out of memory while running a workflow with large data or large amounts of data, you can set Taverna to store data in a database by going to Preferences -> Data and Provenance and removing the tick for “In-memory storage.” Using a database will slightly reduce execution speed, but should consume less memory. Also remember to remove any old runs you are no longer interested in.
User guide: Data and provenance preferences has more details on the data settings in Taverna.
Increase available memory¶
If you need to increase the available memory, for Windows and Linux, edit the taverna.vmoptions (and executeworkflow.vmoptions) file inside Taverna's installation directory (which might require administrator privileges). You can then replace -Xmx700m or -Xmx1g with (for example) -Xmx1500m to use 1.5 GB.
For OS X, modify the Info.plist file and change the same parameter under JavaVM using the Property List Editor. (To modify Info.plist, right-click on Taverna application in Finder and select Show Package Content. Then, open the application bundle to find Contents/Info.plist.)
For platform-neutral Taverna installations not using .vmoptions, you may instead have to modify the taverna.sh/taverna.bat (Windows) and executeworkflow.bat/executeworkflow.sh files.
User guide: Memory allocation has more details on configuring memory allocation.
Taverna Standalone requirements¶
Java and Graphviz are included in the platform-specific downloads but not in the zip downloads.
If you download a platform-neutral Taverna installation, you will need to install Java 7 or higher.
For Oracle's Java 7, you will also need to install the Java Cryptography Extension separately.
For OpenJDK 7, the Java Cryptography Extension is already included. On Linux, apt-get install openjdk-7-jdk, or equivalent, is usually sufficient.
Taverna 2.5 has not been tested with Java 8.
On newer versions of Mac OS X, Taverna Workbench might need the XQuarts shared X11 libraries to be installed.
If you download the platform-neutral Taverna Workbench installation, it includes binaries of the Graphviz tool, which is used for generating workflow diagrams in Taverna. Binaries are included in the bin folder for OS X, Linux x86/x64, and Windows.
In most cases, Taverna should be able to pick up the location of the Graphviz binary dot from the download after you have installed it. If this is not the case, see User guide: Workbench preferences on how to configure the location of dot executable in Taverna Workbench.