SPSCQueue.h VS HPX

A triple buffer is a good choice if all you want is polling the latest data at any given time, and you want to avoid mutexes altogether. If you want each piece of data to be delivered exactly once, you can use a queue (bounded or "unlimited" though the latter doesn't supply backpressure which I hear causes problems). SPSC lock-free bounded queues are dead simple to write, and can be tuned for higher throughput even with contention (https://github.com/rigtorp/SPSCQueue claims to be nice, and I haven't had issues working with it aside from having to peek and pop separately, but it's C++, and not a misuse-proof API since it doesn't use the "handles" idea I talked about, and you can push/read/pop from the wrong thread). If you want the reader to poll/WaitForMultipleObjects until the queue has items, that has to be done separately from the SPSC.

And mutexes make a lot of things easier... and introduces "oops wrong mutex!" (Rust solves it) and deadlock (Rust doesn't solve it).

https://github.com/STEllAR-GROUP/hpx Modern C++ concepts incorporated in a threading library. Lots of useful techniques used in there and we are trying to keep our code base very tidy. Feel free to chime in our libera channel #ste||ar if you have any questions.

https://github.com/STEllAR-GROUP/hpx Welcoming community + we have been part of GSoC for 4-5 years now so feel free to apply there when it opens ;)

You definitely do not want to learn Boost, trust me. The cudatoolkit is fine, HPX is great, so are Dask, and Ray. I do not recommend MPI unless those computers you have use InfiniBand.

I found this interesting library that might fit your goal.

I personally like the ideas in Parallelism v2 TS, which is available in for libstdc++ 11 onwards. The reference implementation is a library named Vc (afaik Vc is the most popular SIMD library for C++), and this has also been implemented in recent versions of HPX.

I am aware of DAGs of task based threading library like Taskflow and HPX however the benefit they have is not obvious to me, as the following sequential section depends on the parallel part being completed fully. If you want to suggest elaboration on the benefits of this approach would be welcome.

Github: https://github.com/STEllAR-GROUP/hpx

We're working on something similar, but based on sender/receiver in HPX (a lightweight threading runtime) and DLA-Future (distributed linear algebra currently based on (HPX) futures; based on sender/receiver in the future). With senders-as-awaitables this would also get you coroutine support for asynchronous MPI calls for free. We don't have that yet, but it's planned. In the meantime libunifex should be able to fill in the gaps.