it would be nice if these benchmark reports actually specified which tasks they passed and which ones they didn't.