There doesn't seem to be a flag for allocating memory. I know that there is this --limitGenomeGenerateRAM however, this seems to be used to constrain how much ram STAR can use...
By default does STAR detect/utilize as much memory as is currently available?
It uses as much memory as it needs, and if that is more than available the job will crash. STAR, pretty much as its only real downside, is quite memory-hungry.
Check out https://github.com/alexdobin/STAR/wiki/FAQ:-Run-time
These numbers generally agree with my experience too. The big ticket item here is the index. If you are working with a large genome, then your index is going to be quite big and all of it gets loaded into memory. On top of that, you will use about 150Mb for every thread.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6