Closed HannesHolste closed 6 years ago
This fix builds on #40.
This makes eigsh find eigenvalues near 1, and just in case, bumps up the max iterations to 5000 it will try before failing.
I have already rerun the benchmarks on the cluster a week ago with this new fix.
Coverage increased (+0.03%) to 87.429% when pulling 6d59299f823f77c5a598d7d534f2ae1d5a229f0d on eigsh-fix-2 into b4f7e9b9e3081d4f5bc5e15cba028c8b98013d26 on master.
Can you delete cell 27 from the notebook? Should make the reading easier ...
@antgonza that ipynb shouldn't have been in this branch (I have a separate analysis branch) – thanks for catching. I removed it.
This fix builds on #40.
This makes eigsh find eigenvalues near 1, and just in case, bumps up the max iterations to 5000 it will try before failing.
I have already rerun the benchmarks on the cluster a week ago with this new fix.