We have unit tests and integration tests, but I would also like to start testing the performance of key operations -- this is often part of "acceptance tests". I think this will require at least a medium-sized test data set, and we should get the average of several runs. The tests themselves should just take a few seconds. What we're trying to catch is changes that make the tests take twice as long as before. And it should also be possible to compare performance of various versions over time, to make sure that we're not getting a lot slower.
We have unit tests and integration tests, but I would also like to start testing the performance of key operations -- this is often part of "acceptance tests". I think this will require at least a medium-sized test data set, and we should get the average of several runs. The tests themselves should just take a few seconds. What we're trying to catch is changes that make the tests take twice as long as before. And it should also be possible to compare performance of various versions over time, to make sure that we're not getting a lot slower.