Expand math_brute_force in the conformance lists to list each subtest.

[anholt]

If you're using a test runner to parse the .csv and shard execution across multiple threads, spawning the test per thread helps balance across CPUs. On my workstation, running wimpy test_bruteforce on a radeon with no args takes 5:21, while with parallelization it completes in 4:21. Given the duration of each subtest (generally 3+ seconds), we don't care about the overhead of spinning up the driver extra times.

[anholt]

List generated using the -p argument.

Karol Herbst has a parallel runner that parses the output of various tests with specific flags and regexes to print and parse their subtests. I would rather we either had this, or a standardized subtest printing in the tests, so that we don't need multiple parsers. I'm happy to pursue either route, but wanted to start with this MR before I go generating and pasting subtest lists for other tests.

[anholt]

List generated using the -p argument.

Karol Herbst has a parallel runner that parses the output of various tests with specific flags and regexes to print and parse their subtests. I would rather we either had this, or a standardized subtest printing in the tests, so that we don't need multiple parsers. I'm happy to pursue either route, but wanted to start with this MR before I go generating and pasting subtest lists for other tests.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}