benchmarking

I just learned about this today, and I thought gcj folks would be
interested:

http://www.stefankrause.net/wp/?p=6

I didn't look at the benchmarks too closely.  It does confirm
something we've known a long time, though, which is that bounds- and
store-checks are slow.  Maybe someone ought to resurrect the tree-ssa
optimization passes for this :)

Tom