Stan Seibert

re: the heterogenously typed mutable containers. The limitation in Numba is not so much the type system as the compilation model that the type system is designed to serve. Numba's...

The cast to float is only going to be a problem for `uint64`, which is how you are avoiding the problem here. The indices are `uint32`, which will cast up...

The two major patches are: * A fix for an upstream linker issue in MCJIT that manifests on AArch64 platforms. * A larger patch for LLVM that enables autovectorization of...

Within Numba, the calls should be as fast as making the calls from C. The cffi-wrapped functions are called directly by C pointer in the generated code. If someone wanted...

If you mean "can the dtype change if axis is specified?", I don't think so. There are still the handful of reduction functions (like `np.mean`) that can return a different...

We tried to design our caching system to be robust against parallel writes. Basically, each process creates the on-disk ELF in a temporary file, then renames it to its final...

Yeah, we knew that file locking is tricky cross-platform, so the current system is intentionally designed to be lock-free.

Nice! We should be able to debug this from here.

This is a good idea. I think the logical way to do this would be to `@overload` the `time.time()` function and call the required C function to get the time....

This is the function in CPython. We can do something simpler, I think: https://github.com/python/cpython/blob/master/Python/pytime.c#L653-L745