The challenges of high performance computing in the future

By -

Despite the efforts made ​​in recent decades by scientists, the human brain remains a great unknown. How it works, why it does so? and, above all, how can be stemmed with kidney diseases, on the other hand, increasingly common as Alzheimer’s, Parkinson’s or even depression?



The European Union this year started an ambitious project aimed precisely to understand the human brain. The call Human Brain Projectis, in fact, one of his two most ambitious scientific projects (along with the research of graphene), endowed with the not inconsiderable sum of 1,190 million euros. The project, still being promoted by the European Commission, is led by the Ecole Polytechnique Federale de Lausanne in Switzerland (country outside the Union) and, in addition, also has the participation of American and Japanese partners.

Combining all the knowledge we now have of this body and rebuild a simulated brain, piece by piece, is the ultimate goal. This will be possible thanks to the use of supercomputing technology, an area in which Spain has a lot to say. Moreover, our country has a large stake in the project and leads the Division of Cellular and Molecular Neuroscience same.

Participation of the BSC-CNS

The work of the Barcelona Supercomputing Center-National Supercomputing Center will be key in the development of the project.“Without the use of supercomputing technologies and compute-intensive,” says Jesús Labarta , director of the Computer Science BCS-CNS and lead researcher on the project: “It would be unthinkable to study the hundreds of thousands of neurons that are to analyze.Supercomputing is essential for simulations and see how they propagate electrical impulses by neurons, how are activated against other … It also helps students of the brain to optimize the code to perform these studies. “

However, not only brings supercomputing capacity and compute-intensive simulation. “The project is also developing advanced computer technologies and researching art programming models that allow these simulations are developed efficiently. What’s more, it also seeks to replicate the performance of the human brain to make better computers, “adds Labarta.

In this sense, explains, “supercomputing has much to learn from this body that is capable of functioning even when some neurons fail. In addition, the brain, with the enormous capacity it has, is able to function with little energy, and this is precisely one of the great challenges of supercomputing. “

The high energy consumption, achieve to reach exascale and dynamic model are, in the words of Labarta, the great challenges of supercomputing in the future. The energy or, rather, the high consumption of it, remains the great workhorse of these machines compute intensive. “Therefore, we are investigating the BSC and participating in the Mont-Blanc project, which involves the design of supercomputers based on low-power chips, such as mobile devices or chips embedded in other devices like televisions ». For now, the project is ongoing. “It will take at least a year to have the prototype,” says the researcher.

As for the exascale challenge model, ie getting the machines run 1,000 times faster than the most powerful supercomputer today, it is a challenge that can be solved “long-term” indicates Labarta.

Finally, when the increased dynamism, i.e supercomputers that are more flexible, malleable and adaptable to changes in the environment and, above all, more fault tolerant, in this will have much to say precisely the Human Brain Project and what we can learn from him about how to overcome their own mistakes in order to survive. “

He have started blogging on Technology and computer at his college time in 2005 and worked with many reputed organization in India. He wrote many guest post for Technology magazine and newspapers worldwide. His writing and passion about Technology make him different from other writers in the global market. He love to write the review and thoughts on any new Technology and invention in current happenings.

Comments are closed.