Commit fd628fa9 authored by Joseph Weston's avatar Joseph Weston
Browse files

increase eigenvalue tolerance by factor 1000

We are typically concerned with features no smaller than 1e-4 - 1e-5
times the bandwidth. With this change we add some tolerance to
error accumulation while not influencing physical results.
parent 10ad6730
...@@ -1000,7 +1000,7 @@ class GreensFunction(BlockResult): ...@@ -1000,7 +1000,7 @@ class GreensFunction(BlockResult):
# We use ‖Σ‖, not ‖Γ‖, for the tolerance as ‖Γ‖~0 when there # We use ‖Σ‖, not ‖Γ‖, for the tolerance as ‖Γ‖~0 when there
# are no open modes. # are no open modes.
eps = ( eps = (
1e3 * np.finfo(gamma.dtype).eps * np.linalg.norm(sigma, np.inf) 1e6 * np.finfo(gamma.dtype).eps * np.linalg.norm(sigma, np.inf)
) )
N = np.sum(np.linalg.eigvalsh(gamma) > eps) N = np.sum(np.linalg.eigvalsh(gamma) > eps)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment