Execute LLVM IR code generated from Python code

Question

Does anyone know if I can execute LLVM IR which is generated from Python code?

Using numba, llvm-lite we can have LLVM IR, but I am not sure if it is executable using clang or not?

I'm guessing that you'll get a better response on the numba mailing list (https://groups.google.com/a/continuum.io/forum/#!forum/numba-users) for this type of question. — JoshAdel, Jun 07 '17 at 20:22

arrowd · Answer 1 · 2017-06-08T06:37:54.817

2

You can compile it with clang, just make sure your IR have .bc or .ll extensions.

But the proper way is to use llc or LLVM API to compile down to native code, and then link it using system linker or LLD.

On linking stage you'd probably need to link some "runtime" libraries, numba ones in this case.

edited Jun 08 '17 at 06:37

answered Jun 07 '17 at 20:45

arrowd

score 0 · Answer 2 · answered Jun 08 '17 at 08:44

0

You can use lli tool which can be found in LLVM's bin folder.

Just do lli file.ll or lli file.bc.

answered Jun 08 '17 at 08:44

2 Answers2