.. _pde_results:

Results
-------

The standard final-results bundle is in
``proj-0/systems/final-results/``:

.. code-block:: console

   $ vmd final.viz.psf final.gro -e final.viz.tcl

Diagnostic-log plots:

.. code-block:: console

   $ htpolynet plots diag --diags diagnostics.log

.. figure:: pics/densification-density.png

   Density vs. time during the densification of the DGEBA/PACM liquid.
   With ``initial_density: 300 kg/m³`` and a single 300 ps NPT segment,
   the system reaches roughly ambient polymer density on the first
   pass.

.. figure:: pics/cure_info.png

   Left: cure conversion vs. wall-clock.  Right: cure iteration index
   vs. wall-clock.  The shape is typical: ~80 % conversion in the
   first half of the cure wall time, the remaining 20 % takes a
   comparable amount of wall time because the late iterations find
   only a handful of bonds each.

For end-to-end traces from ``edr`` files:

.. code-block:: console

   $ htpolynet plots build --proj proj-0 --buildplot t --traces t d p

.. figure:: pics/buildtraces.png

   Top: temperature vs. time across the full build (with cumulative
   bond count overlaid).  Middle: density vs. time.  Bottom: potential
   energy vs. time.

Before and after
^^^^^^^^^^^^^^^^

Snapshots of the initial liquid vs. the cured network, with all bonds
to crosslink sites rendered in licorice.  DGEBA molecules are mauve,
PACM are green:

.. list-table::

    * - .. figure:: pics/dge-pac-liq.png

           System before cure.

      - .. figure:: pics/dge-pac-cured.png

           System after cure.

Profile
^^^^^^^

The end-of-run profile (in ``console.log`` and
``proj-0/profile.json``) shows where the wall time went.  In a typical
DGEBA/PACM run, expect:

* ``setup`` (parameterization of 22 templates) consumes a meaningful
  share — antechamber + tleap dominate this stage;
* ``cure`` iterations themselves are dominated by ``gmx-mdrun`` for the
  per-iteration relax + equilibrate cascades;
* ``capping`` is fast (single iteration over the ~20 leftover oxiranes).

Comparing ``profile.json`` between a run with
``CURE.controls.min_bonds_per_iteration: 1`` and the default
``min_bonds_per_iteration: 10`` is a useful way to see how the knob
trades iteration count for batch size (the iteration counts measured
on this system are tabulated in :ref:`pde_run`).