optimized Icarus Verilog to speed up run time

[Gang Li]

Original Icarus Verilog is based on interpreter virtual machine and run time is slow. We've added a flow to compile the simulation execution to optimized C code, and once compiled into executable, the run time is much faster than original Icarus. We've done some extensive test on many test cases and seen 20-100x speed on designs like video codec, multi-core CPUs, on regular x86 box. Here is some results:

https://github.com/steveicarus/iverilog/discussions/561

I'm wondering if this could be deployed on edaplayground, so we could get more feedback from users for further improvements.