pyextal.jaxGeo.simulate
=======================

.. py:module:: pyextal.jaxGeo.simulate

.. autoapi-nested-parse::

   A memory-safe `jaxEngine.utils.simulate`.



Functions
---------

.. autoapisummary::

   pyextal.jaxGeo.simulate.simulate_safely


Module Contents
---------------

.. py:function:: simulate_safely(params, metadata, config, constants, batch=None, memory_gb=1.5)

   Runs `simulate`, sizing the batch to fit and halving it on OOM.

   The forward model holds a ``(batch, nbeams, nbeams)`` structure matrix, so
   the footprint is ``batch * nbeams**2 * 16`` bytes -- with 565 beams that is
   5.1 MB per pixel, and `simulate`'s default batch of 4096 asks for 20 GB.
   That is what fails after a refinement that itself ran fine, because training
   chunks are far smaller than the simulation batch.

   :param params: Normalised parameters.
   :type params: dict
   :param metadata: From `setup_optimization_params`.
   :type metadata: dict
   :param config: The configuration used.
   :type config: OptConfig
   :param constants: Bloch constants; supplies ``nbeams``.
   :type constants: dict
   :param batch: Pixels per batch. Sized from ``memory_gb`` when None.
   :type batch: int, optional
   :param memory_gb: GB to aim for per batch.
   :type memory_gb: float

   :returns: The simulated intensities.
   :rtype: np.ndarray

   :raises RuntimeError: If even the smallest batch runs out of memory.