This takes the code that was sitting in benchmarks/
already and makes it easier for language-specific
benchmarks to consume. Future PRs will enhance this
so that the language-specific benchmarks can report
metrics back that will be tracked over time in PerfKit.