BY ISABELLE DOM

At Rice, the next breakthrough in cancer therapy, climate modeling or urban design may not come from a lab bench, but from a humming bank of GPUs at the Center for Research Computing. Its new system, RANGE, gives faculty the power to run AI models at a scale once reserved for the world’s biggest research institutions.

RANGE — the Rice AI Networked GPU Engine — is a high-performance computing system built to help faculty tackle major challenges in energy, health and urban innovation. Now, a researcher can screen thousands of new materials before lunch, train a robotic arm overnight or rebuild a 3D medical image in the time it takes to get coffee.

RANGE includes 80 NVIDIA H100 and H200 GPUs — some of the fastest AI processors available today. These processors allow researchers to move huge amounts of information extremely quickly and run advanced AI models in hours or days instead of weeks or months.

More simply, RANGE allows Rice faculty to do research that wouldn’t be possible on standard computers. (Ever try to run a Zoom meeting with one too many programs open?)

By offering researchers access to the kind of computing power typically found only at national labs or major tech companies, RANGE can offer faster discoveries in everything from medical imaging to materials science while bringing new resources and opportunities to Rice and providing hands-on experience for students with advanced AI tools that shape tomorrow’s workforce.

Earlier this year, the CRC invited 11 faculty-led research groups to test and prepare RANGE for campus use. With support from CRC staff, 32 researchers ran more than 9,000 jobs during this pilot phase.

Early projects included:

• Teaching robots to move like humans: Using video clips, researchers trained AI models to mimic natural arm movements for safer, more intuitive robotic motion.
  
• Creating clearer 3D medical images: The Digital Health Initiative tested AI tools that rebuild 3D images from limited data — work that could one day improve the accuracy of radiation therapy.
  
• Discovering new materials faster: One group used AI to explore thousands of possibilities for growing better silicon crystals, which are essential for solar panels and electronics.
  
• Finding new proteins: Two groups are working on finding proteins that are fundamental for designing drugs that can target them for cancer therapies or developing monoclonal antibodies.

The CRC is currently in Phase 2 of its RANGE Early Adopter Program. Each selected awardee received a set allocation of GPU hours to help them test and scale their ideas. Moving forward, the CRC will also host workshops and training sessions to help new users get started, optimize their workflows, and get the best performance from RANGE.

When RANGE comes fully online, it will stand as one of Rice’s most powerful research assets — accelerating discovery across disciplines and strengthening the university’s position as a leader in high-impact, AI-driven science.