A5077055240- Parallel Computing and Optimization Techniques
- Real-Time Systems Scheduling
- Embedded Systems Design Techniques
- Advanced Neural Network Applications
- Distributed and Parallel Computing Systems
- Adversarial Robustness in Machine Learning
- Cloud Computing and Resource Management
ETH Zurich
2023-2025
With the rise of specialized hardware and new programming languages, code optimization has shifted its focus towards promoting data locality. Most production-grade compilers adopt a control-centric mindset - instruction-driven augmented with scalar-based dataflow whereas other approaches provide domain-specific general purpose movement minimization, which can miss important control-flow optimizations. As two representations are not commutable, users must choose one over other. In this paper,...
The same computations are often expressed differently across software projects and programming languages. In particular, how involving loops varies due to the many possibilities permute compose loops. Since each variant may have unique performance properties, automatic approaches loop scheduling must support different optimization recipes. this paper, we propose a priori nest normalization align nests reduce variation before optimization. Specifically, define apply criteria, mapping with...