Exploiting multi-level parallelism in streaming applications for heterogeneous platforms with GPUs

Leiden Repository

Exploiting multi-level parallelism in streaming applications for heterogeneous platforms with GPUs

Title: Exploiting multi-level parallelism in streaming applications for heterogeneous platforms with GPUs
Author: Balevic, Ana
Publisher: Leiden Institute of Advanced Computer Science (LIACS), and Leiden Embedded Research Center, Faculty of Science, Leiden University
Issue Date: 2013-06-26
Keywords: Accelerators
Heterogeneous platforms
Multi-level parallelization
Parallel programming
Polyhedral model
Streaming applications
Abstract: Heterogeneous computing platforms support the traditional types of parallelism, such as e.g., instruction-level, data, task, and pipeline parallelism, and provide the opportunity to exploit a combination of different types of parallelism at different platform levels. The architectural diversity of platform components makes tapping into the platform potential a challenging programming task. This thesis makes an important step in this direction by introducing a novel methodology for automatic generation of structured, multi-level parallel programs from sequential applications. We introduce a novel hierarchical intermediate program representation (HiPRDG) that captures the notions of structure and hierarchy in the polyhedral model used for compile-time program transformation and code generation. Using the HiPRDG as the starting point, we present a novel method for generation of multi-level programs (MLPs) featuring different types of parallelism, such as task, data, and pipeline parallelism. Moreover, we introduce concepts and techniques for data parallelism identification, GPU code generation, and asynchronous data-driven execution on heterogeneous platforms with efficient overlapping of host-accelerator communication and computation. By enabling the modular, hybrid parallelization of program model components via HiPRDG, this thesis opens the door for highly efficient tailor-made parallel program generation and auto-tuning for next generations of multi-level heterogeneous platforms with diverse accelerators.
Description: Promotor: Ed F. Deprettere,Co-promotor: Bart Kienhuis
With summary in Dutch
Faculty: Faculteit der Wiskunde en Natuurwetenschappen
Citation: Balevic, A., 2013, Doctoral thesis, Leiden University
Series/Report no.: ASCI dissertation series;number 283
Handle: http://hdl.handle.net/1887/21017

Files in this item

Description Size View
application/pdf Full text 6.837Mb View/Open
application/pdf Front cover 144.6Kb View/Open
application/pdf Title page_Contents 205.4Kb View/Open
application/pdf Chapter 1 Introduction 1.046Mb View/Open
application/pdf Chapter 2 499.9Kb View/Open
application/pdf Chapter 3 1.778Mb View/Open
application/pdf Chapter 4 1.773Mb View/Open
application/pdf Chapter 5 783.8Kb View/Open
application/pdf Chapter 6 362.2Kb View/Open
application/pdf Chapter 7 Conclusion 233.1Kb View/Open
application/pdf Appendices 515.4Kb View/Open
application/pdf Bibliography 217.4Kb View/Open
application/pdf Index 169.7Kb View/Open
application/pdf Summary in Dutch 169.1Kb View/Open
application/pdf Acknowledgements_Curriculum Vitae 779.0Kb View/Open
application/pdf Propositons 163.3Kb View/Open

This item appears in the following Collection(s)