We need some benchmarks (and ideally an automatic benchmark runner, but that's a different thing). This issue is not to author the benchmarks per se but to make up a candidate list of benchmarks. I think we want a combination of microbenchmarks for particular operations (e.g., creating tracked structs, accessing their fields) and some larger benchmarks, perhaps based on the tutorial project (see #492).
We need some benchmarks (and ideally an automatic benchmark runner, but that's a different thing). This issue is not to author the benchmarks per se but to make up a candidate list of benchmarks. I think we want a combination of microbenchmarks for particular operations (e.g., creating tracked structs, accessing their fields) and some larger benchmarks, perhaps based on the tutorial project (see #492).