Jeffrey Bosboom

David remarked to me that "OpenTuner seems like a very general tool that assumes a very specific usage", and he's right. If someone is using OpenTuner to directly optimize some...

Workaround: run a short tuning session and grep the final configuration JSON.

OpenTuner is not a constraint solver, so those kinds of dependencies between parameter values are not directly expressible. Suppose a hill-climb technique suggests increasing `Mb` by 5 and `Kb` by...

I'm told Una-May knows a solution from the literature.

I don't think so. I mean that the Mario code wants to a) record it used NaiveRepresentation/DurationRepresentation/etc and Progress/ProgressTimesAverageSpeed/etc and b) record the actual progress and elapsed frames for each...

I'd be happy just adding a column for an application-defined JSON string (or whatever the application wants to put in there) on tuning runs and results. Alternately, we could decide...

What seems to be happening is that one of the techniques is producing a configuration with an out-of-range value for that parameter. Then, later, one of the techniques (possibly a...

You shouldn't have any problem running multiple independent tuning runs on multiple machines. At least, when I've done it, I automatically end up with one database file per machine, so...

As a hack I added if self.best_result is not None and self.best_result.time == 0: return True to `convergence_criteria`. We'd want to make the target value a command-line option for use...

I highly recommend trying approach 1 or 3 from [jansel's comment in #106](https://github.com/jansel/opentuner/issues/106#issuecomment-359638307) first. Approach 3 is the clean way to handle this; approach 1 is less "nice" but seems...