Symmetric Transfer in Coroutines

I believe the performance of the coroutine implementation in the benchmark can be significantly improved, to at least match the throughput of [experimental coroutines in Rust](https://github.com/ashvardanian/less_slow.rs/blob/132de76ff7c8a852c2381615ce57f24f5d1706b7/less_slow.rs#L355-L429). The current implementation is relatively minimal, uses `std::suspend_always`, and implements no Symmetric Transfer.

Changing those can help accelerate the benchmarked pipeline. To learn more about coroutine implementations, the [Deciphering C++ Coroutines talk by Andreas Weis from CppCon 2022](https://youtu.be/J7fYddslH0Q?si=f-MHR0-FdtZ9-sHD) is a good start.

Start by locating and running the benchmarked coroutine pipeline, then read its current coroutine implementation and compare it with the linked experimental Rust coroutines. Review the Deciphering C++ Coroutines talk for symmetric transfer and suspend behavior. Done means the C++ benchmark uses the improved coroutine approach and reaches throughput comparable to the stated Rust reference.

Written by the indexing model from the issue text.