Zdravstvyite,
I intend to model a chemical reaction on silica support using a periodic boundary condition approach. The system will contain a transtion metal, Si, C, H, O. The proble is that I have never done it before (all my experience is limited only to molecular-level computations, up to 160 atoms). Could you recommend a fastest software which is using the plane waves approach? After quick search I have identified several packages which could in principle be used for my task:
- VASP (it is installed)
- NWChem (I have an experience in installation, so no problem to get it installed)
- CP2K
- Quantum Espresso
Which one is the fastest and the most reliable? I am going to use PBE functional accomplished with D3 correction. Any suggestions, comments are very appreciated!
reaction os surface
reaction os surface
Кто смел тот и съел
-
Чеширский кот
- Сообщения: 14
- Зарегистрирован: Пт дек 07, 2012 1:45 pm
Re: reaction os surface
Самый скоростной - VASP, единственное ограничение - лицензия.
Для PWscf, Abinit, (возможно и для NWChem) существуют версии с поддержкой GPU, в итоге скорость почти сопоставимая с VASP (но все же чуть меньше) при одинаковом количестве процессоров.
Для PWscf, Abinit, (возможно и для NWChem) существуют версии с поддержкой GPU, в итоге скорость почти сопоставимая с VASP (но все же чуть меньше) при одинаковом количестве процессоров.
Re: reaction os surface
Thank you for reply! Yes, I decided to go for VASP (though I don't like when the software is not free for academic users). We have a licence, no problem. And, from what I also learned, you are right, this is the fastest and the most reiable solid state DFT code. I have also been told that VASP can also be GPU-accelerated. However, it is probably not that strightforward as in case of PWScf.
Кто смел тот и съел
Кто сейчас на конференции
Сейчас этот форум просматривают: нет зарегистрированных пользователей и 25 гостей