Is resource starvation an observable behavior?

[sunfishcode]

For the purposes of the list of all the permitted observable behavior differences, is resource starvation of a thread considered observable behavior?

Are there any guarantees that WebAssembly could make about thread scheduling with respect to resource starvation or fairness?

[jfbastien]

In the C++ standards committee's SG1 concurrency and parallelism group, Torvald Riegel has been specifying things in terms of forward-progress guarantees. I think we want to follow his lead.

See paper N4439 (which will soon be outdated with a newer revision).

[sunfishcode]

The present resolution is that single threaded execution has a forward progress guarantee (though I'm not entirely clear on what that means), and that this will be carried forward to multiple-threaded execution as well. I'm ok with this for now, and even though there may be significant different performance profiles between implementations, I'm ok not adding this to the list for now.