Ilhan Polat
Ilhan Polat
If `N=0` then it will quick exit and `WORK` will not be addressed so I think that is not a problem.
Ah that can be too. Netlib site is down so I can't look at the source code but either it will raise an error or quick exit due to N=0,...
There are two cases in the first one `LWORK=-1` and `N = 0`. In this case, WORK(1) will be set to 0 and calling SGEQRF with this `N=0` and `LWORK...
> If I began squashing, removing, or editing commits, the development history turns into a novel. If you use `fixup` option instead of `squash`, it is as if that commit...
If `info` is not `0` the result should not be trusted in my opinion.
My bad, confused with gesv. Sorry for the noise.
I have been doing the annual linalg check of SciPy overhaul and currently linalg.eig is underway. It seems to me that `uplo` or `lower=True` is a LAPACK plumbing issue rather...
yes indeed and there we start seeing what seems to be a lot of cache misses
Hmm maybe the I should test it with previous versions then. I don't think switching instance is the problem but after switching the blocking or recursion parameters maybe don't match...
Ah :) I know exactly how you feel when I see SciPy versions from 2016. So my presumption was wrong since I was reading [the netlib implementation](http://www.netlib.org/lapack/explore-html/dd/d9a/group__double_g_ecomputational_ga0019443faea08275ca60a734d0593e60.html#ga0019443faea08275ca60a734d0593e60) (line 164) and...