Soroush Ghodrati

Papers

Total Citations

H-Index

About

Soroush Ghodrati is a researcher at the forefront of computer architecture and domain-specific acceleration, with a focus on bridging the gap between specialized hardware and general-purpose programmability. His work centers on developing computational stacks that enable cross-domain acceleration, allowing diverse algorithmic domains to benefit from high-performance hardware without sacrificing expressiveness. Ghodrati’s most cited paper, "A Computational Stack for Cross-Domain Acceleration" (2021, 7 citations), challenges the traditional trade-off between hardware specialization and software flexibility. He proposes a unified framework that supports multiple domains, reducing the need for custom languages and architectures for each application area. This contribution is pivotal for advancing efficient, scalable computing systems, particularly in an era where domain-specific accelerators are increasingly vital for performance gains. Ghodrati’s research has implications for machine learning, data analytics, and scientific computing, where cross-domain efficiency is critical. His work stands out for its practical approach to hardware-software co-design, earning recognition among peers for its potential to reshape how accelerators are built and deployed.