Caleb Winston

Hmm.. it looks like [`PyMPI_reduce_p2p`](https://github.com/mpi4py/mpi4py/blob/98174ae85f7ffaddf252b1e5343095208e29afd0/src/mpi4py/MPI/msgpickle.pxi#L1072-L1110) is what we would want to implement.

Any update on this?

Even if you can address type safety, what about data races? Data races are unsafe in Rust. How would you eliminate that unsafety? I've been trying to figure out type...

I might take back the 2 things I thought were needed for safety. Glium is really, really cool. Honestly can't believe I haven't seen it before. I'll have to look...

By the way, I started experimenting with safety for kernels, programs (not buffers) [here](https://github.com/calebwin/safe_ocl/blob/master/src/lib.rs). This will probably go through multiple make-overs to get the right design of a safe abstraction....

> In theory yes, but in practice it depends on how well methods from BaseMatrix and BaseMatrix map to the idea of parallelizing your logic and type constraints imposed by...

Seconding this - this would be a really useful feature!

I'm having the same issue and it would be great to get some insight on this.

Just to be clear - Emu is as "pure Rust" as WebGPU is. It's just a thin abstraction over WebGPU (providing things like a global device pool, global JIT kernel...

> any plans to support any other intermediate languages (SYCL, RLSL)? Personally, I don't think SYCL or RLSL are actually very useful as abstractions. I feel that something like [Halide](https://people.csail.mit.edu/jrk/halide-pldi13.pdf)...