Of late it seems that everybody has been throwing around their own
little homegrown benchmarks to support their points.  But many people
frequently point out that these benchmarks are flawed on one way or another.

I suggest that we add a benchmark/ subdirectory and create a canonical
suite of benchmarks that exercise things well (and hopefully fully).
Then we can all post relative times for runs on this benchmark suite,
and we will know exactly what is being tested and how valid it is.

Matt