Construct non-negative lambdas and budget constraints

Constructs the CVX objective function.

Convert input fraction into number of partitions

Estimate the cost of an experiment. Right now this is input_frac/machines

Returns a list of covariance matrices given expt design features

Compute the features for a given point. Point is expected to be [input_frac, machines]

Enumerate all the training points given the params for experiment design

Predict running time for given input fraction, number of machines.

Predict running time for a batch of input sizes, machines. Input test_data should be a list where every element is (input_fraction,

Run experiment design. Returns a list of configurations and their scores

Initiliaze the Predictor with some training data The training data should be a list of [mcs, input_fraction, time]

Create an experiment design instance. :param self: The object being created :type self: ExperimentDesign :param part