Frequently Asked Questions

This is a list of Frequently Asked Questions about GYRE. Suggestions for new entries are always most welcome!

How Do I…

…obtain the GYRE source code?

See the Download section.

…compile GYRE?

See the Compile section.

…compile a statically linked version of GYRE?

Set the STATIC environment variable to yes prior to compilation. Note that this currently only works on Linux platforms, and when then CRMATH environment variable is set to no.

…run GYRE on multiple cores?

GYRE can take advantage of multiple processors in a shared-memory (multicore) computer through its use of OpenMP. This functionality should be enabled by default, but you can nevertheless force it by setting the OMP environment variable to yes prior to compilation. Then, set the OMP_NUM_THREADS environment variable to the number of threads you want to use.

…run GYRE on a cluster?

GYRE can take advantage of multiple nodes in a computer cluster through its use of MPI. To enable this functionality, set the MPI environment variable to yes prior to compilation. Note that you’ll need to use a version of the MESA Software Development Kit built with MPI support for your cluster’s specific networking technology (e.g., Infiniband); contact the GYRE team for more details.

…pronounce GYRE?

With a soft ‘g’ rhyming with ‘wire’, like this.

…cite GYRE?

See the Citing GYRE section.

…get support for a problem I’m having?

Post a message to one of the GYRE discussion forums.

…contact the GYRE team?

Send an email to Rich Townsend.

…access the documentation for older releases?

Click on the ‘v:…’ dropdown menu at the bottom of the left-hand panel. Note that this menu is only available when viewing the documentation on Read the Docs; it isn’t available in a local build within the $GYRE_DIR/doc/sphinx directory. Also, the legacy documentation is a work-in-progress, and remains incomplete.

What Does…

…’Failed during deflate narrowout-of-domain frequency’ mean?

This is an indication that GYRE’s root solver wandered out of bounds when trying to find a complex root of the discriminant function. Try running with a different choice of diff_scheme parameter (MAGNUS_GL2 seems to be the most robust), and/or using contour method instead (see the Non-Adiabatic Oscillations chapter).

Why Does…

…the error ‘Illegal Instruction’ arise on MacOS when running with large grid sizes?

This behavior is typically caused by overflow of the OpenMP stack (for more info see here). Try setting the OMP_STACKSIZE environment variable to 500K or 1M.