Use single-threaded executor in benchmarks by default

[alice-i-cecile]

What problem does this solve or what need does it fill?

As investigated in #9835, the overhead of multithreading can make benchmarks noisy in ways that are irrelevant to what we're actually testing.

What solution would you like?

Implement this solution executed in the linked PR on our other benchmarks where relevant. Except when we're explicitly testing system parallelism, we probably don't want it.

[hymm]

This is probably true for testing too. Though in that case it's more for execution speed. It might be worth adding a single_threaded constructor for this to make it explicit.

[Telov]

Why does it have D-Good-First-Issue ?

[alice-i-cecile]

The changes themselves are pretty mechanical, there's a working example of how this might be done, and the stakes are very low :)