An analysis and experimental approach to teaching data prefetching on CMP

ABSTRACT

At our university, since the Spring of 2005, we have been teaching a first course about data cache on CMP from computer architecture. We have accomplished several goals. The most important of which is the analysis and experimental approach for pull-based data prefetching and push-based data prefetching on CMP. The pedagogical style embodied in this course fosters a good understanding of the relationship between memory wall and data prefetching for the students early in their experience.

REFERENCES

• 1: John L. Hennessy , David A. Patterson, Computer Architecture, Fourth Edition: A Quantitative Approach, Morgan Kaufmann Publishers Inc., San Francisco, CA, 2006
• 2: Wenmin Zheng, et al. 1998. Computer Architecture, Second Edition, Tsinghua University Press.
• 3: Guoliang Chen, et al. 2002. Parallel Computer Architecture, Academic Education Press.
• 4: Zhimin Gu, Xianhe Sun. 2008.Parallel Computer Architecture and Scalable Computing, Tsinghua University Press. (in edit and publish)
• 5: Surendra Byna, Yong Chen, Xian-He Sun. 2007. A Taxonomy of Data Prefetching Mechanisms, Technical Report N<u>o</u>IIT/CS-SCS07--01, May.
• 6: Seungryul Choi , Nicholas Kohout , Sumit Pamnani , Dongkeun Kim , Donald Yeung, A general framework for prefetch scheduling in linked data structures and its application to multi-chain prefetching, ACM Transactions on Computer Systems (TOCS), v.22 n.2, p.214-280, May 2004 [doi>10.1145/986533.986536]
• 7: Jason Mars, Daniel Williams, Dan Upton, Sudeep Ghosh, Kim Hazelwood. 2008. A Reactive Unobtrusive Prefetcher for Multicore and Manycore Architectures. SHCMP 2008
• 8: Intel® VTune#8482; Performance Analyzer for Linux. http://www.intel.com/support/performancetools/vtune/linux
• 9: Intel® 64 and IA-32 Architectures Software Developer's Manual Volume 3B: System Programming Guide, Part 2.www.intel.com

INDEX TERMS

Primary Classification:
  C. Computer Systems Organization
  C.1 PROCESSOR ARCHITECTURES

General Terms:
Algorithms, Experimentation, Measurement

Keywords:
data prefetching, data push