Benchmark bonding curve functionality

[ethanfrey]

Follow up from #185

Let's get some idea how fast these algorithms are to see if we need to optimize the implementations. Or if they are good enough.

• Benchmark the various Curve implementations (native rust) ns/run
• Test with gas meter the buy/burn with various Curve implementations (at least a fast and two slow variants from the benchmark). See if wasm gas usage is proportional to the benchmark times.

[ethanfrey]

Setting up benchmarking infrastructure in general. Both timing (as native rust) as well as gas cost (as wasm in cosmwasm-vm) would be interesting. Not sure if native rust benchmarking is really useful, as the wasm gas costs are what matter and are deterministic (only need to run once, no average).

@maurolacy just want to ping this for a tooling task that may be a useful way to learn the stack

[ethanfrey]

Interesting idea, but I think no one is using cw20-bonding anyway, certainly not me.