Ethan Wu
Papers
1
Total Citations
5
H-Index
1
About
Ethan Wu is a pioneering computer architect whose research focuses on domain-specific accelerators for machine learning and robotics. His most notable contribution is the NeCTAr (Near-Cache Transformer Accelerator), a 16nm heterogeneous multicore RISC-V SoC that efficiently handles both sparse and dense machine learning kernels. This groundbreaking work, detailed in his 2024 paper "NeCTAr and RASoC," introduces a novel architecture combining near-core and near-memory accelerators to optimize matrix-vector multiplications, achieving 109 GOPS at 400MHz while operating at just 0.85V. The prototype chip demonstrates how specialized hardware can dramatically improve energy efficiency for language model inference and robotic applications. Though early in his career, Wu's work has already garnered attention with 5 citations, and his innovative approach to integrating heterogeneous accelerators on a single chip positions him as an emerging leader in the field of efficient AI hardware. His research directly addresses the growing need for specialized processors that can handle the computational demands of modern AI workloads while maintaining power efficiency, making his contributions particularly relevant for edge computing and autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1