Add large matrix inversion

[oxinabox]

This PR adds large matrix inversion.

I feel it might fit here under the category of "A place to experiment with fast linear algebra routines written in Julia" It is hard to find a home for this code.

But it is really quiet useful and fast.

It was mostly written by Mahdi Jamei. (Who I do not think is on Github). I just did some cleanup, and optimisations and am handling releasing it somewhere open source

[codecov-io]

Codecov Report

Merging #46 into master will decrease coverage by 21.88%. The diff coverage is 45.45%.

@@             Coverage Diff             @@
##           master      #46       +/-   ##
===========================================
- Coverage   72.77%   50.88%   -21.89%     
===========================================
  Files          10       11        +1     
  Lines        1381     1411       +30     
===========================================
- Hits         1005      718      -287     
- Misses        376      693      +317

Impacted Files	Coverage Δ
src/inv.jl	45.45% <45.45%> (ø)
src/eigenGeneral.jl	34.78% <0%> (-44.1%)	:arrow_down:
src/householder.jl	39.74% <0%> (-42.54%)	:arrow_down:
src/cholesky.jl	67.64% <0%> (-32.36%)	:arrow_down:
src/qr.jl	72.54% <0%> (-25.5%)	:arrow_down:
src/tridiag.jl	53.84% <0%> (-23.08%)	:arrow_down:
src/lapack.jl	28.04% <0%> (-22.77%)	:arrow_down:
src/eigenSelfAdjoint.jl	69.44% <0%> (-16.95%)	:arrow_down:
src/svd.jl	59.1% <0%> (-16.6%)	:arrow_down:
src/rectfullpacked.jl	78.57% <0%> (-7.54%)	:arrow_down:
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update ccb13bc...c5e5a59. Read the comment docs.

[oxinabox]

@andreasnoack can you take a look at this?

[andreasnoack]

I have quite a bit on my plate for the time being but will review this eventually.

[oxinabox]

bump

[ViralBShah]

May I suggest we merge this so long as the tests are passing? @oxinabox In case you would be up for refreshing, that is.

[andreasnoack]

Do we know anything about the error properties of this algorithm?

[oxinabox]

I don't. I know in practice it is not great, but it is usable. I didn't do the math for working this out, i just tidied up the code and made the PR