Resolves NIT-3023

• Adds more WASM instructions to be benchmarked
• Adds cargo check CI step related to stylus_benchmark
• Renames exec_program to launch_program_thread
• Uses random inputs in some instructions. Some instructions, such as popcnt, have different performances for different inputs. We try to avoid input performance bias by calling the same instruction multiple times, and using random inputs. It was observed that multiple benchmark runs of the same scenario have similar running times.
• Now --help CLI argument shows the possible scenarios that can be benchmarked