Open syclik opened 9 years ago
I was looking at this issue and looking through the code for inverse_spd and log_determinant_spd. Both use Eigen's LDLT and other Eigen functions. Is this still an issue? Or, are you looking to further optimize inverse_spd and log_determinant_spd?
@syclik or @mrbrubake: Could one of you verify this issue should still be outstanding.
And if it is, will whoever works on it please get rid of all those redundant //
comment markers and unused includes and move the TODO
items to an issue if they're still relevant.
This should still be outstanding but it's primarily a speed issue and (to a lesser extent) a robustness/stability thing. It would be an easy fix for someone, I suspect @rtrangucci or anyone else with a familiarity with LDLT could do it in an hour or less, as it looks like there are already tests for it.
From @mbrubake on February 18, 2014 22:44
Currently still uses inverse_spd/log_determinant_spd which may be less than optimal for large systems.
Copied from original issue: stan-dev/stan#574