Start benchmarking

[milktrader]

Besides the test suite, we should begin benchmarking that not only do we get the correct answer, but we get it fast.

[milktrader]

I've included the test/data/benchmark.csv file, which includes over 15,850 rows of data. Not sure if this is big enough for benchmarking. I think that taking simple returns of the Close for example would give a nice array to time functions on.

[milktrader]

We should follow the layout that DataFrames has for this, which is similar to how unit testing is done.

[rob-luke]

Is there a way to automatically run @time or other benchmarking tools automatically via travis or some other CI on original and changed code? So you can see if a pull request will provide performance improvements.

[rob-luke]

The Benchmark package looks nice. But I see no easy way to compare against different git commits.

[andreasnoack]

Such functionality is actively being worked on right now so hopefully it'll be availiable within a couple of months.

[rob-luke]

Thanks. Is there a conversation somewhere I can follow?