What would be the impact of such a coroutine emulation if the threading is used to leverage multi-core hardware for high performance computing such as done by Atlas [1], OpenBLAS [2] or MKL [3]? These libraries are tuned to maximize CPU cache hits. It seems to me that executing each thread task sequentially using coroutines would probably break such optimizations.

[1] http://math-atlas.sourceforge.net/ [2] http://www.openblas.net/ [3] http://software.intel.com/en-us/intel-mkl

From the sounds of it, you'd want to split up your workload and run each subset in a separate zerovm

That's correct. And that's how most of parallel processing frameworks do it anyway.