Reduce more linear algebra to multiplication

[wbhart]

Now that we have parallel matrix multiplication, we should ensure more linear algebra is reduced to multiplication to make use of it.

[kartikv]

Is there brief documentation on how to do parallelism in Flint? I'm happy to have the sparse matrix multiplies be parallel, and I'm guessing dense Gaussian elimination-based operations could also be nicely parallelized (lots of inner loops).

[wbhart]

There's nothing very comprehensive yet. You can take a look in code_conventions.txt where there is something. There's also some documentation in flint.rst and thread_pool.rst. But the best way to do it at present beyond that is to look at some examples.

We do need to write something more comprehensive on this.