Regression for 128-bit arithmetic support

[stevefan1999-personal]

ZLUDA's LLVM does everything correctly and there's no issue. NVIDIA's CUDA also returns an error on cuModuleLoadDataEx.
It's rust-cuda's LLVM that fails to include __multi3

Originally posted by @vosen in #562

[LegNeato]

I think this should be fixed by #278, can you try that one?