Portuguese 
English 
Spanish 
2 min of reading 
0 comments 
Human Brain-Inspired Computing Can Transform How Drugs Are Discovered. A New Supercomputer Promises to Accelerate This Process with Unprecedented Energy Efficiency
Leipzig University has partnered with the technology company SpiNNcloud to receive the world’s largest supercomputer with a human brain-inspired architecture.
The system, equipped with around 650,000 cores, will primarily be used for research on small molecules for personalized medicine treatments.
Additionally, the machine will be applied in protein simulation tasks, essential for the advancement of new drugs.
-
A computer creates a thermoelectric generator with an unusual shape, more precise heat flow, and 8 times higher efficiency than the conventional rectangular version.
-
While billion-dollar aircraft carriers remain the ultimate symbol of naval power, drones, fast boats, and mines in the Strait of Hormuz show that controlling trade routes has become a more complex challenge, even for the world’s largest Navy today in narrow strategic areas.
-
ChatGPT becomes a central piece in a lawsuit against OpenAI after an attack at a US university, and the case raises an uncomfortable question about how far artificial intelligence can be held responsible.
-
A study by Nature Geoscience already has a date for the end of oxygen in Earth’s atmosphere; the research indicates that the increase in solar luminosity will reduce the CO₂ available for photosynthesis, leading to an abrupt drop in the gas in about 1 billion years.
The project represents a milestone in the integration of artificial intelligence, high-performance computing, and biotechnology.
System with 4320 Chips and Billions of Simulated Neurons
The supercomputer features 4320 chips from the second generation of SpNNaker hardware. This technology was originally developed by Steve Furber, the same engineer responsible for the ARM architecture.
The system configuration allows for the simulation of up to 10.5 billion neurons, enabling applications in AI, HPC, and other scientific uses.
According to Christian Mayr, co-founder of SpiNNcloud, the system enables massive digital screening of molecules. He claims that the architecture allows testing billions of molecules “in silico,” that is, via computational simulation.
“The SpiNNcloud Server System can screen 20 billion molecules in less than an hour, which is 100 times faster than a thousand traditional CPU cores,” Mayr highlighted.
Energy Efficiency as a Differentiator
The system’s design combines 48 SpiNNaker2 chips per server board. Each chip has 152 ARM-based cores and features specific accelerators. This configuration allows for high performance with lower energy consumption.
According to Hector Gonzalez, CEO of SpiNNcloud, reduced energy consumption is one of the most important advantages. He states that the company’s systems are up to 18 times more efficient than conventional GPUs.
“Our architecture allows for deploying algorithms that leverage dynamic scarcity and extreme parallelism,” Gonzalez explained. He also noted that the adoption of this technology is growing in leading institutions across Europe and the United States.
A New Path for AI
Most importantly, SpiNNcloud is enabling a different approach in the field of artificial intelligence. Unlike traditional models, which use the same fixed set of neural connections, the new system works with extreme dynamic scarcity.
This means that, depending on the input, only a subset of neural pathways is activated.
As a result, AI can learn more efficiently and consume less energy, potentially representing a significant shift in how AI models are built.
Therefore, the initiative not only reinforces the importance of biomimetic computing but also opens new paths for the future of medicine and artificial intelligence. The installation of the supercomputer in Leipzig may mark the beginning of a new era in drug discovery.
-
-
-
4 people reacted to this.