Optimizing Sparse Matrix Exponential Computation in Julia

General Usage
,
1

I’m trying to compute the sparse matrix exponential of large matrices (at least 10,000 x 10,000) using Julia packages like Expokit, FastExp, and ExpmV. However, I’m finding that the computation time is significantly longer compared to using Python’s scipy.sparse.linalg.expm_multiply.

For example, I’m using the function expmv(1.0, M, rho_vectorized), where M is a sparse matrix.

What methods or strategies can I use to speed up the computation of the sparse matrix exponential in Julia? Any advice or tips would be greatly appreciated!

2

Have you tried using ExponentialUtilities.jl? That’s the standard package for this.

1 Like
3

Thank you for your response. I found a bug in my code that caused it to run slower than Python. However, when I use ExponentialUtilities.jl, it is very quick, but the result is incorrect. Maybe I use a wrong function. I will check it again. Thx.