NumPy aware dynamic Python compiler using LLVM
Numba is a translator of Python byte-code to LLVM IR which can then be used to create machine code. It is a dynamic compiler which is used intentionally in places where you might otherwise want to write C/C++, Fortran, or Cython code.
Example:
@numba.vectorize
def sinc(x):
if x == 0:
return 0.0
else:
return sin(pi*x)/(pi*x)
This will compile to machine code using LLVM and build a NumPy ufunc that uses the corresponding loop for its code.
@teoliphant @jriehl @jaberg @markflorisson88 @sklam
Sign up for mailing list: https://groups.google.com/a/continuum.io/forum/#!forum/numba-users