David Nellans
Papers
1
Total Citations
3
H-Index
1
About
David Nellans is a leading researcher in computer architecture, with a primary focus on the intersection of memory systems, GPU computing, and emerging workloads like machine learning and distributed reinforcement learning. His work has been instrumental in understanding how to design efficient, scalable hardware for modern AI training. Notably, his highly cited paper "The Architectural Implications of Distributed Reinforcement Learning on CPU-GPU Systems" (2020, 3 citations) provides a critical analysis of the performance bottlenecks and power challenges that arise when scaling deep RL across heterogeneous CPU-GPU platforms. This research has helped shape the design of next-generation accelerators by revealing how architectural choices—from memory hierarchy to interconnect design—directly impact the scalability and efficiency of distributed AI training. Nellans’ contributions are essential for researchers and engineers building the hardware foundations for tomorrow’s intelligent systems, bridging the gap between algorithmic advances and practical, high-performance computing.
Research Focus
Key Achievements
Top Papers
- 1