[Benchmark] conjugate gradients

[willwng]

A bril benchmark for using conjugate gradients to solve Ax=b. Currently the file takes in n, which is the size of the matrix and A is diagonal with incrementing values. However, the code implements dense matrix multiplication and can support any arbitrary A (it is somewhat tricky to randomly generate spd matrices).

fyi: the vim syntax highlighting also sometimes stopped working when functions had a specified return type

[sampsyo]

Really cool; nice work!

There is a minor conflict on the docs page; could you please resolve that so we can merge?

[sampsyo]

Thanks!! Now that the CI has run, it has also revealed that the benchmark has a few type errors. (You can run the type checker yourself if you want.) In particular, it looks like some places there is a ptr<int> where it should probably be a ptr<float>.

[willwng]

Thanks!! Now that the CI has run, it has also revealed that the benchmark has a few type errors. (You can run the type checker yourself if you want.) In particular, it looks like some places there is a ptr<int> where it should probably be a ptr<float>.

Oops, didn't realize there was a type checker! The github workflow integration is really neat. I think I fixed the type errors and added an input for the size of the matrix.

[sampsyo]

Looks great; thanks!