Output "terminated with early" during CREST. Problem?

[codecollider]

Hi,

I´m currently running some CREST runs with a non covalently bound complex made up of two molecules (currently both neutral)

As the molecules separate during a default CREST run I´m using the option NCI. I didn´t find any issues when not using NCI, but with NCI I´m getting messages like "*MTD 5 terminated with early ... 2 min, 32.042 sec" in the output. The calculation is still proceeding though and I wonder if this is a problem.

This is the total output until now:

`
╔════════════════════════════════════════════╗
║ ___ ___ ___ ___ _____ ║
║ / | _ \ / | _| ║
║ | (| / |_ \ | | ║
║ _||_|_/ || ║
║ ║
║ Conformer-Rotamer Ensemble Sampling Tool ║
║ based on the xTB methods ║
║ ║
╚════════════════════════════════════════════╝
Version 3.0.2, Fri Jul 19 02:13:53 UTC 2024
commit (f1b6928) compiled by 'runner@fv-az1211-722'

Cite work conducted with this code as

• P.Pracht, F.Bohle, S.Grimme, PCCP, 2020, 22, 7169-7192.
• S.Grimme, JCTC, 2019, 15, 2847-2862.
• P.Pracht, S.Grimme, C.Bannwarth, F.Bohle, S.Ehlert,
G.Feldmann, J.Gorges, M.Müller, T.Neudecker, C.Plett,
S.Spicher, P.Steinbach, P.Wesołowski, F.Zeller,
J. Chem. Phys., 2024, 160, 114110.

for works involving QCG cite

• S.Spicher, C.Plett, P.Pracht, A.Hansen, S.Grimme,
JCTC, 2022, 18 (5), 3174-3189.
• C.Plett, S. Grimme,
Angew. Chem. Int. Ed. 2023, 62, e202214477.

for works involving MECP screening cite

• P.Pracht, C.Bannwarth, JCTC, 2022, 18 (10), 6370-6385.

Original code
P.Pracht, S.Grimme, Universität Bonn, MCTC
with help from (alphabetical order):
C.Bannwarth, F.Bohle, S.Ehlert, G.Feldmann, J.Gorges,
S.Grimme, C.Plett, P.Pracht, S.Spicher, P.Steinbach,
P.Wesolowski, F.Zeller

Online documentation is available at
https://crest-lab.github.io/crest-docs/

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Lesser General Public License (LGPL) for more details.

Command line input:
$ crest xtbopt.xyz --nci

--nci : Special NCI mode for non-covalently bound complexes or clusters.

constraint: wall_fermi atoms: 221/all
radii(AA)= 10.60436 9.98546 12.23104
kb*T= 0.00094 exp= 50.00000
Automatically generated ellipsoide potential for NCI mode:

Generating logfermi wall potential for the system.
wscal= 1.000 wpad= 0.000 atoms: all
constraint: wall_fermi atoms: 221/all
radii(AA)= 10.60436 9.98546 12.23104
kb*T= 0.00094 exp= 50.00000

Setting up backup calculator ... done.

---

Calculation info

User-defined calculation level:
: xTB calculation via tblite lib
: GFN2-xTB level
: Molecular charge : 0
: Fermi temperature : 300.00000
: Accuracy : 1.00000
: max SCC cycles : 500
: Read dipoles? : yes
: Weight : 1.00000

User-defined constraints:
constraint: wall_fermi atoms: 221/all
radii(AA)= 10.60436 9.98546 12.23104
kbT= 0.00094 exp= 50.00000
constraint: wall_fermi atoms: 221/all
radii(AA)= 10.60436 9.98546 12.23104
kbT= 0.00094 exp= 50.00000

---

Initial Geometry Optimization

Geometry successfully optimized.

      ┍━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┑
      │              CREST iMTD-GC SAMPLING             │
      ┕━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┙

Input structure:
################################ taken out by me ########################################

Generating MTD length from a flexibility measure

System flexiblity is set to 1.0 for NCI mode
flexibility measure : 1.000

! WARNING: the estimated MTD time exceeds 200.0 ps.
! Because the estimate is uncertain, the program restricts
! this to 200.0 ps and continues. The user may
! re-run crest with manual setting by '-mdlen ' and
! check the results carefully.

t(MTD) / ps : 200.0
Σ(t(MTD)) / ps : 1200.0 (6 MTDs)

---

Starting trial MTD to test settings

Trial MTD 1 runtime (1.0 ps) ... 14 min, 35.591 sec
Estimated runtime for one MTD (200.0 ps) on a single thread: 48 h 38 min 38 sec
Estimated runtime for a batch of 6 MTDs on 32 threads: 9 h 43 min 44 sec

---

** N E W I T E R A T I O N C Y C L E **

---

---

Meta-Dynamics Iteration 1

list of applied metadynamics Vbias parameters:
$metadyn 0.22100 1.000
$metadyn 0.11050 1.000
$metadyn 0.22100 0.500
$metadyn 0.11050 0.500
$metadyn 0.22100 0.250
$metadyn 0.11050 0.250

::::::::::::: starting MTD 5 :::::::::::::
| MD simulation time : 200.0 ps |
| target T : 300.0 K |
| timestep dt : 5.0 fs |
| dump interval(trj) : 100.0 fs |
| SHAKE algorithm : true (all bonds) |
| dump interval(Vbias) : 1.00 ps |
| Vbias prefactor (k) : 0.2210 Eh |
| Vbias exponent (α) : 0.2500 bohr⁻² |
::::::::::::: starting MTD 6 :::::::::::::
| MD simulation time : 200.0 ps |
| target T : 300.0 K |
| timestep dt : 5.0 fs |
| dump interval(trj) : 100.0 fs |
| SHAKE algorithm : true (all bonds) |
| dump interval(Vbias) : 1.00 ps |
| Vbias prefactor (k) : 0.1105 Eh |
| Vbias exponent (α) : 0.2500 bohr⁻² |
::::::::::::: starting MTD 3 :::::::::::::
| MD simulation time : 200.0 ps |
| target T : 300.0 K |
| timestep dt : 5.0 fs |
| dump interval(trj) : 100.0 fs |
| SHAKE algorithm : true (all bonds) |
| dump interval(Vbias) : 1.00 ps |
| Vbias prefactor (k) : 0.2210 Eh |
| Vbias exponent (α) : 0.5000 bohr⁻² |
::::::::::::: starting MTD 4 :::::::::::::
| MD simulation time : 200.0 ps |
| target T : 300.0 K |
| timestep dt : 5.0 fs |
| dump interval(trj) : 100.0 fs |
| SHAKE algorithm : true (all bonds) |
| dump interval(Vbias) : 1.00 ps |
| Vbias prefactor (k) : 0.1105 Eh |
| Vbias exponent (α) : 0.5000 bohr⁻² |
::::::::::::: starting MTD 1 :::::::::::::
| MD simulation time : 200.0 ps |
| target T : 300.0 K |
| timestep dt : 5.0 fs |
| dump interval(trj) : 100.0 fs |
| SHAKE algorithm : true (all bonds) |
| dump interval(Vbias) : 1.00 ps |
| Vbias prefactor (k) : 0.2210 Eh |
| Vbias exponent (α) : 1.0000 bohr⁻² |
::::::::::::: starting MTD 2 :::::::::::::
| MD simulation time : 200.0 ps |
| target T : 300.0 K |
| timestep dt : 5.0 fs |
| dump interval(trj) : 100.0 fs |
| SHAKE algorithm : true (all bonds) |
| dump interval(Vbias) : 1.00 ps |
| Vbias prefactor (k) : 0.1105 Eh |
| Vbias exponent (α) : 1.0000 bohr⁻² |
*MTD 5 terminated with early ... 2 min, 32.042 sec
*MTD 6 terminated with early ... 42 min, 20.939 sec
*MTD 1 terminated with early ... 51 min, 9.004 sec
*MTD 2 terminated with early ... 16 min, 0.166 sec
*MTD 3 terminated with early ... 37 min, 16.056 sec
*MTD 4 terminated with early ... 37 min, 31.512 sec

======================================
| Multilevel Ensemble Optimization |

Optimizing all 2072 structures from file "crest_dynamics.trj" ...

crude pre-optimization

Optimization engine: ANCOPT
Hessian update type: BFGS
E/G convergence criteria: 0.500E-03 Eh, 0.100E-01 Eh/a0
|>0.0% |>7.5% |>15.0% |>22.5% |>30.0% |>37.5% |>45.0% |>52.5% |>60.0% |>67.5% |>75.0% |>82.5% |>90.0% |>97.5% |>100.0%
done.

2072 of 2072 structures successfully optimized (100.0% success)
Total runtime for 2072 optimizations:

• wall-time: 0 d, 3 h, 5 min, 37.280 sec
• cpu-time: 1 d, 15 h, 54 min, 21.684 sec
• ratio c/w: 12.899 speedup

Corresponding to approximately 5.375 sec per processed structure

CREGEN> running RMSDs ... done.
CREGEN> E lowest : -353.56801
586 structures remain within 12.00 kcal/mol window

optimization with tight thresholds

Optimization engine: ANCOPT
Hessian update type: BFGS
E/G convergence criteria: 0.100E-05 Eh, 0.800E-03 Eh/a0
|>0.2% |>10.1% |>20.1% |>30.0% |>40.1% |>50.0% |>60.1% |>70.1% |>80.0% |>90.1%`

I can email the input structure as well if it helps. I deleted it from the input above as this is possibly a hot topic.

Regards,
Marco

[pprcht]

No, it's not a problem unless the MDs terminate immediately. Since the metadynamics modify the potential over time, sometimes the simulation becomes unfeasible and terminates earlier than the targeted runtime. The reason why some terminate earlier than others is that different metadynamics strengths are applied.

[codecollider]

Thank you!