Without staging optimization, RBNF seems to be much slower than original CPython parser(with CPython parser, import sklearn costs 2~3 seconds, and RBNF costs almost 10 seconds).
Staging optimization makes RBNF very fast, however a huge startup overhead makes it useless.
To release a version with friendly compiler services, we try to figure a way to perform JIT techniques on bottom-up parser generator and reduce the overhead of wrapping Python interfaces.
https://github.com/thautwarm/rbnfrbnf
Without staging optimization, RBNF seems to be much slower than original CPython parser(with CPython parser,
import sklearncosts 2~3 seconds, and RBNF costs almost 10 seconds).Staging optimization makes RBNF very fast, however a huge startup overhead makes it useless.
To release a version with friendly compiler services, we try to figure a way to perform JIT techniques on bottom-up parser generator and reduce the overhead of wrapping Python interfaces.
https://github.com/thautwarm/rbnfrbnf