Interesting enough, this naive implementation is ~30% more space efficient and yields ~10% faster CPU. I can't explain the former, but the latter is possibly due to better register allocation for Op.decode and Op.step (2 bytes each), as opposed to 16 bytes.
closes https://github.com/fengb/fundude/issues/27
Interesting enough, this naive implementation is ~30% more space efficient and yields ~10% faster CPU. I can't explain the former, but the latter is possibly due to better register allocation for Op.decode and Op.step (2 bytes each), as opposed to 16 bytes.