Improved BSGS to Decrease Number of Rotations and Address #2162

[lawrencekhlim]

Currently having trouble running and testing at the moment. Might like for someone else to help test.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[lawrencekhlim]

A few updates: the code improves the 512x512 matrix multiplication test case, but I haven't tested it too carefully for correctness (though I think it should be correct). Adding a few more test cases for correctness may help. All of the test cases now pass though.

Unfortunately, I'm not sure if I'm doing something wrong, but when I try to build locally using the Docker image and bazel, it takes more than an hour to build, so it's not very convenient for me to test this much more. @j2kun, maybe you could look at this a bit more and try it with your bicyclic matrix multiplications?

[j2kun]

Thank you!! This looks much simpler. I feel it may be the right time to finish onboarding fuzz tests #2196

[j2kun]

I think what I'll do with this PR is cherry-pick the commit, rebase over the fuzz testing work, and then fix any build issues and test it more thoroughly. Then I'll squash the commits together and keep @lawrencekhlim as a coauthor to ensure he gets credit :)

[lawrencekhlim]

Sounds good to me!

[j2kun]

Moving to #2392

Thanks!!!