The current code for indexing into bucket arrays is quite tricky and unsafe,
partly because it has to keep manually assuring the compiler that a bucket
index is always less than 21.
By encapsulating that knowledge in a 21-value enum, we can make the code
clearer and safer, without giving up performance.
Having a dedicated `BucketIndex` type could also help with further cleanups of
`VecCache` indexing.
This replaces the single Vec allocation with a series of progressively
larger buckets. With the cfg for parallel enabled but with -Zthreads=1,
this looks like a slight regression in i-count and cycle counts (<0.1%).
With the parallel frontend at -Zthreads=4, this is an improvement (-5%
wall-time from 5.788 to 5.4688 on libcore) than our current Lock-based
approach, likely due to reducing the bouncing of the cache line holding
the lock. At -Zthreads=32 it's a huge improvement (-46%: 8.829 -> 4.7319
seconds).
Zalathar