MC measurements

Detailed Description

Abstraction for Monte Carlo measurements.

A measurement is any object that, given the current sign of the Markov chain, contributes to one or more running estimators. This group provides the type-erased wrappers used by triqs::mc_tools::mc_generic to store heterogeneous measurements together:

• triqs::mc_tools::measure wraps a user object satisfying triqs::mc_tools::MCMeasure and bookkeeps the number of measurements, optional timing and an optional report() callback.
• triqs::mc_tools::measure_set is the container of registered measurements; it forwards accumulate and collect_results to every measurement it owns.
• triqs::mc_tools::measure_aux holds auxiliary callables that are run on every cycle but do not themselves accumulate an estimator (e.g. live diagnostics or periodic state dumps).

Final reduction across MPI ranks is delegated to the user code inside collect_results, which receives the communicator.

Classes
class	triqs::mc_tools::measure< MCSignType >
	Type erasure class for MC measures. More...
class	triqs::mc_tools::measure_aux
	Type erasure class for auxiliary MC measures. More...
class	triqs::mc_tools::measure_set< MCSignType >
	Set of MC measures. More...