Open
Description
We haven't set up much standards for 'good' benchmarks here.
My ideas are
- Running benchmarks across every major release of SymPy, and check if there is any failing one
- Check the minimum, and maximum time for the benchmarks in the range.
And running in travis before merging.
We may need some scripts for this