High performance computing


Theoretical chemistry and Modelling


A coupled theory/experimental approach

 

The evolution of research tools in the fields of

  • Modelling and representation of complex chemical systems
  • structural analysis, characterisation and data processing

requires the acquisition and mutualisation of costly computing resources and access to an efficient network infrastructure, responding to the constraints inherent to calculations performed in the field of chemistry.

Theoretical chemistry and modelling are essential for the study of matter, either as discrete molecules or as solids. Theoretical studies contribute to the structural determination of chemical systems and enhance the understanding of the electronic properties that determine the physical and chemical characteristics measured experimentally. These studies are also particularly useful for understanding chemistry transformations and are crucial for the representation of chemical systems. 

 

Extensive and diversified computational resources for performance and reliability.

 

Theoretical chemistry is a long way from achieving its goal, which is to connect the electronic properties of groups of atoms to measurable macroscopic quantities. This is particularly important for nanosystems whose properties vary with the size. Modelling is a real challenge and requires important computational resources. Moreover, this computational technology must be diversified as the methodological approaches are quite varied.
In other words, the system requirements to conduct the necessary studies vary from one subject to another. Some calculations depend on quick access to the storage space, some require enormous quantities of random-access memory, and others need data transfer between nodes in low latency. Computational resources that can simultaneously serve the needs of all these types of calculations can be difficult to find and yet not very profitable, because some characteristics will only be used on an occasional basis. It is recognize by the international community that diversification of the means of calculation is essential for this field. 


These high-performance computational resources are useful in the fields of : 

  • Energy 
    •  New energy sources: hydrogen generation, methanol
    • Storage: lithium-ion batteries, supercapacitors
    • Materials for and conversion devices
    • Fuel cells, photovoltaic cells, light emitting diodes

     

  • Environment
    •  Active principles and vectorisation: new molecules, nano-encapsulation
    • Diagnosis assistance: medical imaging, biosensors, biochips
    • Materials for bio-applications: biomimetic materials, bio-active surfaces

     

  • Health
    •  Pollution detection and treatment: chemical sensors, selective adsorbents
    • Valorisation of agro biological resources: agro-fuels
    • Green chemistry: synthesis with atoms economy, homogeneous and heterogeneous catalysis

 

 

Ressources

 

Members of the ICGM information systems team

  • Fabrice Boyrie (IR2)
  • Cyril Bourgogne (IR2)
  • Mourad Guermache (AI)
  • Sylvain Pallares (IE2)
  • Pierre Sans (IE1)

To be able to handle emerging research problems as well as accommodate new users, the ICGM has increased its computational power each year, leading to a total of 5000 cores. 



Some ICGM high-performance computational resources :

  • 174 nodes based on a bi-XEON X5560 (or equivalent), 8 cores, 24 Go RAM/nodes.
  • 102 nodes based on a bi-Xeon E52670v1 (or equivalent), 16 cores, 64 Go RAM/nodes.
  • 28 nodes à based on a bi-Xeon E52697v3 (or equivalent), 28 cores, 128 Go RAM/nodes.
Most of the machines are inter-connected by "infiniband" low latency networks.


Some of the available software :

  • Gaussian 2003 and 2009
  • Vasp v 5.4 (solid-state quantum chemistry)
  • Wien 2k (solid-state quantum chemistry)
  • Materials Studio (solid-state quantum chemistry)
  • Mctdh
  • Demon, Gamess, Crystal

Institut Charles Gerhardt Montpellier - Direction

UMR 5253 - CNRS/UM/ENSCM
  • Université de Montpellier
  • Place Eugène Bataillon
  • CC 1700 - Bâtiment 17 -1er étage
  • Tel: +33 (0)4 67 14 93 50
  • Email: direction@icgm.fr
Top