Skip to Main Content
 

Global Search Box

 
 
 
 

ETD Abstract Container

Abstract Header

Integrated compiler optimizations for tensor contractions

Gao, Xiaoyang

Abstract Details

2008, Doctor of Philosophy, Ohio State University, Computer and Information Science.
This dissertation addresses several performance optimization issues in the context of the Tensor Contraction Engine (TCE), a domain-specific compiler to synthesize parallel, out-of-core programs for a class of scientific computations encountered in computational chemistry and physics. The domain of our focus is electronic structure calculations, where many computationally intensive components are expressible as a set of tensor contractions. These scientific applications are extremely compute-intensive and consume significant computer resources at national supercomputer centers. The manual development of high-performance parallel programs for them is usually very tedious and time consuming. The TCE system is targeted at reducing the burden on application scientists, by having them specify computations in a high-level form, from which efficient parallel programs are automatically synthesized. The goal of this research is to develop an optimization framework to derive high-performance implementations for a set of given tensor contractions. In particular, the issues investigated include: 1) Development of an efficient in-memory parallel algorithm for a tensor contraction. 2) Design of a performance-model driven framework for a parallel out-of-core tensor contraction. 3) Design of an integrated optimization framework, which incorporating multiple high-level program transformation techniques to improve locality and parallelism for a set of tensor contractions. This subject includes the following three topics: 3.1) Modeling interactions between different transformations and assess their impact on the overall performance, 3.2) Using a search-based strategy to explore the combined space of transformations, and provide efficient pruning methods to reduce the search space, 3.3) Using performance-driven models to identify the best combination and generate high-performance code. 4) Incorporation of domain-specific optimizations, such as sparse and symmetry. We have implemented and evaluated most of the proposed algorithms and optimizing strategies. Experimental results show that they work effectively in practice.
Ponnuswamy Sadayappan (Advisor)
162 p.

Recommended Citations

Citations

  • Gao, X. (2008). Integrated compiler optimizations for tensor contractions [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1198874631

    APA Style (7th edition)

  • Gao, Xiaoyang. Integrated compiler optimizations for tensor contractions. 2008. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1198874631.

    MLA Style (8th edition)

  • Gao, Xiaoyang. "Integrated compiler optimizations for tensor contractions." Doctoral dissertation, Ohio State University, 2008. http://rave.ohiolink.edu/etdc/view?acc_num=osu1198874631

    Chicago Manual of Style (17th edition)