Improve interface for running benchmarks #362
Description
The current dialog for submitting a benchmark run request is very basic:
The list of functions and benchmarks to choose from are just basic dropdown lists which makes it hard to find the specific function you want when there are a lot of published functions.
Selecting the benchmark and task is probably not going to be as much of an issue since we are unlikely to have more than a few benchmarks to pick from, for now.
Ideally we could make a nice function selection selection component we can use in multiple contexts. We have a similar use case on the garden page: #338