Back to Table of contents

Primeur weekly 2017-10-02

Special

2017 - Another year on the Road to Exascale: HPC applications towards exascale ...

Focus

CEA, ETP4HPC and European Commission working hard together to deliver commercial exascale by 2023 ...

Exascale supercomputing

Cartography of the cosmos ...

Quantum computing

A quantum computer to tackle fundamental science problems ...

Quantum communications bend to our needs ...

With new Microsoft breakthroughs, general purpose quantum computing moves closer to reality ...

Focus on Europe

SHAPE Sixth Call for Applications now open ...

EPCC to announce best use of ARCHER competition winners ...

Middleware

Bright Computing announces integration with IBM Power systems ...

US-based CryoEM company, SingleParticle.com, partners with Bright ...

Hardware

Cray takes aim at fast-growing HPC storage market with completion of strategic transaction of Seagate's ClusterStor storage business ...

Penguin Computing announces NVIDIA Tesla V100-based servers to drive Deep Learning and Artificial Intelligence ...

Penguin Computing announces Altus servers with AMD EPYC 7000 series high performance processors ...

Supermicro introduces broadest portfolio of GPU optimized systems for NVIDIA Tesla V100 GPUs ...

New type of supercomputer could be based on 'magic dust' combination of light and matter ...

Applications

Antonio Peña is the IEEE-CS TCHPC 2017 Award Winner for Excellence for Early Career Researchers in HPC ...

University of Texas at El Paso researchers use supercomputers to understand the fundamentals of artificial photosynthesis ...

Turbocharging engine design ...

The sublime challenge of jet noise ...

Mapping black hole collisions gives astronomers - and hitchhikers - a new guide ...

George Washington-led consortium and FDA release new specifications to advance genomic data analysis ...

Artificial intelligence for obtaining chemical fingerprints ...

IBM Research and UC San Diego collaborate to advance the use of Artificial Intelligence for Healthy Living ...

Dresden researchers have pioneered a brain-network bio-inspired algorithm to predict new therapeutic targets of approved drugs ...

The Cloud

IBM unveils a new high-powered analytics system for fast access to data science ...

HPE chooses Mellanox Spectrum to power StoreFabric M-series switches ...

University of Texas at El Paso researchers use supercomputers to understand the fundamentals of artificial photosynthesis


The structure of the multichromophoric heptad molecule in solution. The molecule has been shown to act as a light harvesting system. Researchers optimized its structure using density functional theory calculations at the Texas Advanced Computing Center.
25 Sep 2017 El Paso - Solar power is booming worldwide, offering an alternative to fossil fuels and mitigating some of the environmental costs of coal and gas burning. However, efficient photovoltaics are still relatively expensive for full-scale, worldwide deployment. In the search for cheap and sustainable materials that can harvest solar power, scientists frequently look to proven solar power generators - plants, algae and bacteria - to develop man-made sustainable sources of energy.

"Scientist have been trying to design efficient, durable devices that can functionally mimic light harvesting processes as they happen in nature", stated Rajendra Zope, a professor of physics at the University of Texas El Paso (UTEP).

Breakthroughs in nanotechnology and nanostructured materials have led to artificial photosynthetic systems that can convert light into usable energy.

In the 1990s, researchers developed a molecule called PCBM - Phenyl-C61-butyric acid methyl ester - a hollow cage of carbon atoms derived from a 'buckyball', that could transform sunlight into electrical energy. However a fully-functional, scalable and stable form of organic artificial photosynthesis system has yet to emerge.

For the development of functional materials, it is necessary to understand the photoconversion process taking place inside the material. In this area, supercomputers play a critical role by allowing researchers to study materials' properties, such as their quantum electronic structure, and by providing insights into how, where and when electrons move through the active material using simulations.

One of the challenges in simulating complex systems like artificial photosynthetic molecules is the size of these entities, which researchers try to model from the ground up - also known as 'ab-initio'. Another challenge lies in trying to pinpoint the charge transfer excited states in detail.

For several years, Rajendra Zope and his colleagues in the UTEP Electronic Structure Lab have used the Stampede and Lonestar supercomputers at the Texas Advanced Computing Center (TACC) - some of the most powerful in the world - to model the electronic structure of artificial photosynthetic systems.

Writing in the journal Chemical Physics , they described calculations related to a photosynthetic, multichormophoric antenna system comprising of 421 atoms that is able to absorb light from many different wavelengths.

"Our simulations are done to understand the energy level ordering of the system", he stated. "The ordering is very important to see if the charge transfer exciton will deliver the current or not."

In addition to studying the ordering, they explored how the inclusion of various ligands - molecules bound to a metal atom - changed the dynamics of the system, making it more or less effective at transforming light into energy. They also predicted the charge transfer energy of the molecular complex in various configurations.

The work is supported by grants from the Department of Energy and the National Science Foundation.

Before Rajendra Zope and his collaborators can run their simulations, they must first develop the computer code capable of mathematically modelling the quantum energetics of systems as big as the ones they study.

The software they develop and maintain for this purpose is known as NRLMOL - the Naval Research Laboratory Molecular Orbital Library. A massively parallel code for electronic structure calculations on large molecules and clusters, it is used by more than a dozen groups around the world to study not only light harvesting systems, but also molecular magnets and transitional metal systems - both promising systems for next-generation nano-electronics.

Creating effective algorithms entails using a large number of computer processing cores concurrently while minimizing the memory usage on every processor.

"There are very few codes that scale as well as our code in terms of the number of cores it can use", Rajendra Zope stated. "For the excited state calculations of the light harvesting hexad containing 421 atoms, we can use up to 1,000 processors with 70 percent parallel efficiency. This is quite good for density functional electronic structure codes like NRLMOL that use Gaussian basis sets."

Rajendra Zope and his team have used TACC systems since 2009. They were first able to access TACC systems through the University of Texas Research Cyberinfrastructure (UTRC) programme, which makes TACC's computing resources, expertise and training available to researchers within the University of Texas Systems' 14 institutions.

They also access TACC systems through the Extreme Science and Engineering Discovery Environment (XSEDE), which allocates time to researchers on national supercomputing resources. In addition to TACC supercomputers, Rajendra Zope's team uses systems at the U.S. Department of Energy's National Energy Research Scientific Computing Center.

Light harvesting research has been advancing in recent years, but has yet to reach the level of widespread commercial use. However, Rajendra Zope believes that in a decade, many of the fundamental questions about light harvesting nanostructures will have been worked out and the area will blossom.

"It would have a tremendous effect", Rajendra Zope stated. "Per person, energy use is continuously growing and the population is growing, so the energy demand is significant. Once we figure out the mechanism to improve efficiency, then all we are doing is harnessing the light. If we realize this, we will have solved the energy problem in a nice, clean manner."
Source: University of Texas at Austin, Texas Advanced Computing Center - TACC

Back to Table of contents

Primeur weekly 2017-10-02

Special

2017 - Another year on the Road to Exascale: HPC applications towards exascale ...

Focus

CEA, ETP4HPC and European Commission working hard together to deliver commercial exascale by 2023 ...

Exascale supercomputing

Cartography of the cosmos ...

Quantum computing

A quantum computer to tackle fundamental science problems ...

Quantum communications bend to our needs ...

With new Microsoft breakthroughs, general purpose quantum computing moves closer to reality ...

Focus on Europe

SHAPE Sixth Call for Applications now open ...

EPCC to announce best use of ARCHER competition winners ...

Middleware

Bright Computing announces integration with IBM Power systems ...

US-based CryoEM company, SingleParticle.com, partners with Bright ...

Hardware

Cray takes aim at fast-growing HPC storage market with completion of strategic transaction of Seagate's ClusterStor storage business ...

Penguin Computing announces NVIDIA Tesla V100-based servers to drive Deep Learning and Artificial Intelligence ...

Penguin Computing announces Altus servers with AMD EPYC 7000 series high performance processors ...

Supermicro introduces broadest portfolio of GPU optimized systems for NVIDIA Tesla V100 GPUs ...

New type of supercomputer could be based on 'magic dust' combination of light and matter ...

Applications

Antonio Peña is the IEEE-CS TCHPC 2017 Award Winner for Excellence for Early Career Researchers in HPC ...

University of Texas at El Paso researchers use supercomputers to understand the fundamentals of artificial photosynthesis ...

Turbocharging engine design ...

The sublime challenge of jet noise ...

Mapping black hole collisions gives astronomers - and hitchhikers - a new guide ...

George Washington-led consortium and FDA release new specifications to advance genomic data analysis ...

Artificial intelligence for obtaining chemical fingerprints ...

IBM Research and UC San Diego collaborate to advance the use of Artificial Intelligence for Healthy Living ...

Dresden researchers have pioneered a brain-network bio-inspired algorithm to predict new therapeutic targets of approved drugs ...

The Cloud

IBM unveils a new high-powered analytics system for fast access to data science ...

HPE chooses Mellanox Spectrum to power StoreFabric M-series switches ...