3 ms·

What is meant by “quickening”?

• rurban 6y ago
  The paper is gone. Maybe someone has a link. Looks like inline caching of some specific ops with certain types. Some arith and call ops are optimized this way. It also looks like dynasm was later replaced by the much slower llvm, so it needs heavy warmup.
  • kzrdude 6y ago
    I wonder if it's similar to Python 3.x's opcache for Load global and load attr? It would make a lot of sense to generalize those.