Comments (6)
Can you try 'atom_style atomic'?
from mace.
Can you try 'atom_style atomic'?
Thanks, I tried and It kinda works. I forgot to update the question.
If I understand correctly, atomic is the style to be used because it is useless to explicitly define bonds and angle between atoms in molecules using this kind of potentials. Maybe I am wrong but if MACE mp0 is trained on potential coming DFT calculations, so an ab initio method, after a number of iterations, water molecules should "form" only because of the potential, without the need of explicitly fixing distances, bonds, etc.
I tried the following two scripts and it works, the problem is that it is apparently too slow and it segfaults for lack of memory I think, but I'm just trying on a small laptop with 16 GB of RAM...
from mace.
Does it work on your laptop for a smaller system? Try the smallest that is reasonable.
from mace.
Does it work on your laptop for a smaller system? Try the smallest that is reasonable.
I tried with a 64 molecules systems, so small enough I think but I have still some segfaults problems.
This is the initial configuration, it is a water system that should be equilibrated
This is the error:
This it the lammps input script:
I tried with both domain decomposition and with no domain decomposition.
Can you try and see if I'm doing something wrong? Thank you.
from mace.
I've tried your example and reproduced similar errors. As best I can tell, the main issue is that you are using a gigantic timestep (1 ps). Try something closer to 1 fs.
from mace.
I've tried your example and reproduced similar errors. As best I can tell, the main issue is that you are using a gigantic timestep (1 ps). Try something closer to 1 fs.
Thanks, that was the problem. With a timestep of 0.001 (1 fs) it seems to work.
Another question:
I'm running it on a cluster with four Intel(R) Xeon(R) Gold 6252N CPU @ 2.30GHz cpus. I found that the simulation still crashes if I run it with more than one processor, for example with mpirun -np 2 lmp
but I have to experiment more.
The best performances are obtained with running a job with just one mpi process but using a large number of threads (96 for example setting the env variable with export OMP_NUM_THREADS=96
) without domain decomposition, just following the lammps mace documentation. With this setting it takes 25 seconds to compute 1 timestep for a system of 125 water molecules starting from an equilibrated configuration using the mace mp0 L1 model. Is it normal to get this results?
By the way the next task is trying to compile the GPU version to run on my modest Nvidia RTX 3060Ti 8 GB on my Arch Linux system :-)
from mace.
Related Issues (20)
- Deployment of the MACE Model Without Using ASE HOT 5
- Error while using multihead interface HOT 2
- Segmentation fault during LAMMPS-MACE-CPU run HOT 1
- Problems with parallelization on CPU (Using LAMMPS) HOT 1
- Improve plotting script
- Automatically add parameters to optimizer
- Change stress to virials internally
- ASE documentation update? HOT 4
- Move to ruff for linting and formating? HOT 3
- Tripeptides Data for MACE-OFF Model
- Bug on loading finetuning model HOT 2
- Remove `e3nn` version pin HOT 2
- models created from multihead fine tuning don't work in lammps HOT 6
- universal loss (at least in multi-head-interface branch) does not support per-config weight
- cannot turn off multihead finetuning in multi-head-interface branch HOT 17
- Colab Tutorial Link Does not work HOT 2
- help message for `--pair_repulsion` is wrong HOT 1
- How to use multi-GPUs training with PBS system HOT 20
- [feature suggest] active learning in lammps
- I got Warning pytorch not support instance-level annotations. HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from mace.