Improve falcon-bench

Things to add:
- Routing benchmark
- "falcon-ext" for all frameworks
- Additional frameworks
- wheezy.web benchmarks
- pyramid benchmark
- Django-rest
- morepath
- WebCore
- Benchmarks exercising various req properties
- Dump data as JSON
- Render body as JSON across different frameworks (vs. plaintext blob)

Things to clean up:
- CLI
- Separate flags for test vs. framework
- Improve code organization

See also:
- https://bitbucket.org/akorn/helloworld/src
- http://www.techempower.com/benchmarks/#section=data-r9&hw=peak&test=json&l=1kw
- https://klen.github.io/py-frameworks-bench/ (But note that the async frameworks tend to win out vs. WSGI frameworks, this should be remedied once we introduce twisted, aiohttp, etc. support)