Large reworking and performance increase on spars solver

[zinkkkk]

Not really an 'issue' but i have been playing around with the spars solver files and managed to get some pretty notable performance improvements mostly by removing the linked list matrix format/container. As far as i can see, the solving portion of this program has remained untouched from its first release in 2004 so any changes to it should be pretty thoroughly reviewed and evaluated but there seems to be a lot of room for improvement. Really i don't even know much about programming this is basically the first thing i've ever done and i also am not familiar with matrix math either so i'm not super confident in the code i have made.

This site confuses me and i genuinely can't work out how i'm supposed to correctly do pushes/commits/pulls and have it send only the bits i want to the right place/branch/whatever. Here is the new/updated stuff i added to my own fork f3dbaca
For example, how would i pull just this commit over? It seems to make push all of my commits at once? Do i need a separate personal branch for every little commit i want to pull separately? Urgh.

Anyway i would like to hear what other people think of these potential updates as in larger femm problems i often find it is able to solve twice as fast or more while small and simple things are roughly the same as previously.

[reos-rcrozier]

I just noticed this issue, might get a chance to test next month, but not before