Skip to content

Efficient implementation of a null-space Rayleigh Quotient Iteration in Jax with sparse matrices. Works on the GPU!

License

Notifications You must be signed in to change notification settings

choltz95/null-space-rayleigh-quotient-iteration

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Null-space-Rayleigh-Quotient-Iteration

Implementation of a null-space Rayleigh Quotient Iteration algorithm in Jax with sparse matrices. Works on the GPU! Could be much faster if we use a better way to solve the linear subproblem.

Solves a problem that looks like:

equation

It can be solved as a generalized eigenvalue problem of PAP, with P the projection onto the subspace orthogonal to v. One issue is that if A is sparse, PAP could be dense and we could run into memory or runtime problems with big matrices. This implementation favors matrix-vector multiplications- we never compute or store PAP & should be able to scale to some pretty big graphs.

About

Efficient implementation of a null-space Rayleigh Quotient Iteration in Jax with sparse matrices. Works on the GPU!

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages