Nvidia's supercomputer toolset: Opening up workings of the human heart in real time

Nvidia and the Barcelona Supercomputing Center are working on building a better cardiac computational model.

nvidia-barcelona-supercomputing-center-heart-simulation.jpg

Nvidia IndeX software will provide in-situ visualizations of the electromechanical simulations produced by the Alya Red project.

Image: Barcelona Supercomputing Center

Physicist Mariano Vázquez is coordinator for the Alya Red project at the Barcelona Supercomputing Center. His role involves translating the human heartbeat into lines of code that supercomputers located in the XIX century chapel of the Polytechnic University of Catalonia can transform into a simulation.

That's not an easy task because the heart "is a multiphysical, very complex system", he says. In fact, simply simulating the way electrical impulses make the heart muscles contract to pump blood takes four hours' work by between 256 and 512 processors.

Almost all the world's fastest supercomputers run Linux

498 out of 500 of the world's speediest supercomputers now use Linux.

Read More

However, a joint research project with graphics chip firm Nvidia may now ease that process, producing faster results and immediate in-situ visualization.

The project will involve the use of Nvidia's IndeX high-performance computing visualization tool, which can work with the tremendous computing capabilities that GPU-accelerated clusters and HPC systems feature today, Vázquez says.

The software enables real-time, high-quality rendering of large-scale data at any dataset resolution and combines it with novel in-situ technologies. For example, it allows technicians "to change simulation parameters, such as electrical impulses, while it's running". That ability could significantly improve the process of placing a pacemaker, he adds.

Doctors are more cautious about the project's immediate application. Primary-care physician Dr Xavier Peral thinks that for predictive purposes, simulations might be practical. "But at the clinical level, coronary angiography is more effective," he adds.

The Alya Red computational biomechanics project, which started in 2007, aims at contributing to a better understanding of biological systems. It has grown remarkably in recent times.

Vázquez acknowledges that images such as CT and MRI scans provided by the Spanish Centro Nacional de Investigaciones Cardiovasculares Carlos III, based in Madrid, have been key "to a better understanding of the uses for the simulation, to improve the clinical and medical component of the research and to validate the work".

Now, Barcelona Supercomputing Center wants to take the project a step further and is starting to work with companies in the biomedical sector, especially those producing medical devices, such as pacemakers, stents, and valves, and those that produce antiarrhythmic drugs.

The ability to simulate large-scale complex problems, combined with a quick visual inspection of the massive computational data produced in an interactive environment, can accelerate the arrival of more personalized drugs and even new healing therapies.

Furthermore, it could potentially cut the time and money spent in research and reduce the extent of animal testing.

Barcelona Supercomputing Center is also creating a spinoff to develop an interface that makes Alya Red easier to use.

However, the Alya Red project is still in its early stages. Mariano Vazquez's dream is to be able to simulate the entire human body and interconnect all its biological systems, but that goal would require a lot of computing power and a huge amount of medical data, as well as visualization tools that allow researchers to interpret it.

This is where the new Marenostrum4 supercomputer may come in, with its performance capacity of 13.7 petaflops, or 13,677 trillion operations per second. It's due to be operational on July 1.

Read more about supercomputing

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All