Capping memory usage?

[Aaron K]

Chris,

I'm interested in finding a way to cap the memory usage in the incremental and global solvers to prevent out of memory situations for larger alignments. This would of course involve caching to disk to accomplish.

I'm going to be looking into this more in the next week or so, and I wanted to see if you'd considered any approaches to this end yet? At initial glance I thought the inclusion of rocksdb was part of this, but it appears that only helps with initial feature detection and matching.

Any insight or thoughts appreciated,

Aaron