Running the Build

From inside the working directory containing 4-dfda-fde-epoxy-thermoset.yaml:

$ htpolynet run -diag diagnostics.log 4-dfda-fde-epoxy-thermoset.yaml &> console.log &

Because the chemistry, CURE controls, precure cascade, and postcure cascade are identical to example 3, the stage layout and the diagnostic-log shape are also the same. Refer back to tutorial 3 for the per-stage narrative; below are the few items that are specific to this build.

Setup

setup parameterizes the same kind of template set as example 3 — FDE and DFA primitives, the four primary-to-secondary cure products, the eight secondary-to-tertiary products, the cap variants, and the diastereomer pool from FDE’s stereocenters. Furan ring perception adds a small overhead compared to the all-carbocycle PACM/DGEBA case, but antechamber and tleap still complete cleanly thanks to the atom-mapped SMILES path described in monomers.

Densification, precure, CURE, postcure

These mirror example 3 exactly — same 300 kg/m³ initial density, same 300 ps NPT densification, same 0.5 nm starting search radius, same 0.95 desired conversion (target 380 of 400 bonds), same 0.85 late threshold, same drag/relax/equilibrate cascades. Expect a similar shape: early iterations grab dozens of bonds; the long tail picks up ones and twos with the search radius growing into the drag regime.

The CURE.controls.min_bonds_per_iteration knob applies the same way here as in example 3. The default of 10 was chosen using measurements on the DGEBA/PACM system (see the table at Running the Build); the same chemistry runs through this system so the sweet spot is likely similar. To explore, drop the knob to 1 and compare profile.json between runs to see the cost of the un-batched regime.

Final

After postcure, htpolynet writes final.gro, final.top, final.tpx, final.grx, the VMD viz pair, and the profile report into proj-N/systems/final-results/. The next page describes what to look at.