Non-transparent debugging for software-pipelined loops

ABSTRACT

This paper tackles the problem of providing correct information about program variable values in a software-pipelined loop through a non-transparent debugging approach. Since modern processors provide instruction level parallelism, software pipelining techniques have been developed to achieve better performances, especially in the context of embedded systems. Indeed, the effectiveness of software pipelining on such systems has been demonstrated both theoretically and experimentally. As it overlaps iterations and reorders statements, it also makes standard debugging information irrelevant. Hence debugging a loop which has been software-pipelined becomes very difficult. In this paper, we propose a solution relying on selected information to be generated by the compiler and an algorithm for the debugger not to mislead the user.

REFERENCES

• 1: A WORKGROUP OF THE FREE STANDARDS GROUP. The DWARF Debugging Standard. http://dwarf.freestandards.org/.
• 2: ACE ASSOCIATED COMPILER EXPERTS BV. CoSy Compilers, Overview of Construction and Operation. White paper, 24 Apr. 2003.
• 3: ADL-TABATABAI, A.-R. Source Level Debugging of Globally Optimized Code. PhD thesis, Carnegie Mellon University, Pittsburgh PA 15213-3891, June 1996.
• 4: Andrew W. Appel , Maia Ginsburg, Modern Compiler Implementation in C, Press Syndicate of the University of Cambridge, 1998
• 5: Valérie Bertin , Jean-Marc Daveau , Philippe Guillaume , Thierry Lepley , Denis Pilat , Claire Richard , Miguel Santana , Thomas Thery, FlexCC2: An Optimizing Retargetable C Compiler for DSP Processors, Proceedings of the Second International Conference on Embedded Software, p.382-398, October 07-09, 2002
• 6: Gary Brooks , Gilbert J. Hansen , Steve Simmons, A new approach to debugging optimized code, Proceedings of the ACM SIGPLAN 1992 conference on Programming language design and implementation, p.1-11, June 15-19, 1992, San Francisco, California, United States
• 7: Max Copperman , Charles E. McDowell, A further note on Hennessy's “Symbolic debugging of optimized code”, ACM Transactions on Programming Languages and Systems (TOPLAS), v.15 n.2, p.357-365, April 1993 [doi>10.1145/169701.214526]
• 8: D. S. Coutant , S. Meloy , M. Ruscetta, DOC: a practical approach to source-level debugging of globally optimized code, Proceedings of the ACM SIGPLAN 1988 conference on Programming Language design and Implementation, p.125-134, June 20-24, 1988, Atlanta, Georgia, United States
• 9: D. M. Dhamdhere , K. V. Sankaranarayanan, Dynamic currency determination in optimized programs, ACM Transactions on Programming Languages and Systems (TOPLAS), v.20 n.6, p.1111-1130, Nov. 1998 [doi>10.1145/295656.295657]
• 10: ELMS, K. Debugging optimised code using function interpretation. In Automated and Algorithmic Debugging (1997), pp. 27--36.
• 11: GOUGH, K., LEDERMANN, J., AND ELMS, K. Interpretive debugging of optimised code. Proceedings of ACSC-17, Christchurch (1994).
• 12: John Hennessy, Symbolic Debugging of Optimized Code, ACM Transactions on Programming Languages and Systems (TOPLAS), v.4 n.3, p.323-344, July 1982 [doi>10.1145/357172.357173]
• 13: Clara Ines Jaramillo , Mary Lou Soffa , Rajiv Gupta, Source-level debugging techniques and tools for optimized code, University of Pittsburgh, Pittsburgh, PA, 2000
• 14: Pierre G. Paulin , Miguel Santana, FlexWare: A Retargetable Embedded-Software Development Environment, IEEE Design & Test, v.19 n.4, p.59-69, July 2002 [doi>10.1109/MDT.2002.1018134]
• 15: B. Ramakrishna Rau, Iterative modulo scheduling: an algorithm for software pipelining loops, Proceedings of the 27th annual international symposium on Microarchitecture, p.63-74, November 30-December 02, 1994, San Jose, California, United States [doi>10.1145/192724.192731]
• 16: STALLMAN, R. M. Using and Porting the GNU Compiler Collection, For GCC Version 2.95. Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA, Tel: [phone], USA, 1999.
• 17: STALLMAN, R. M., PESCH, R., SHEBS, S., ET AL. Debugging with GDB: The GNU Source-Level Debugger. 2002.
• 18: TICE, C., AND GRAHAM, S. Key Instructions: Solving the Code Location Problem for Optimized Code. Compaq, Systems Research Center, 2000.
• 19: Caroline Tice , Susan L. Graham, OPTVIEW: a new approach for examining optimized code, Proceedings of the 1998 ACM SIGPLAN-SIGSOFT workshop on Program analysis for software tools and engineering, p.19-26, June 16-16, 1998, Montreal, Quebec, Canada
• 20: Caroline M. Tice, Non-Transparent Debugging of Optimized Code, University of California at Berkeley, Berkeley, CA, 2000
• 21: David Wall , Amitabh Srivastava , Fred Templin, A note on Hennessy's “symbolic debugging of optimized code”, ACM Transactions on Programming Languages and Systems (TOPLAS), v.7 n.1, p.176-181, Jan. 1985 [doi>10.1145/2363.215005]
• 22: Roland Wismüller, Debugging of globally optimized programs using data flow analysis, ACM SIGPLAN Notices, v.29 n.6, p.278-289, June 1994
• 23: WU, L.-C., AND MEI W. WHU, W. A new breakpoint implementation scheme for debugging globally optimised code. Urbana 51 (1998), 61801.
• 24: Le-Chun Wu , Rajiv Mirani , Harish Patil , Bruce Olsen , Wen-mei W. Hwu, A new framework for debugging globally optimized code, Proceedings of the ACM SIGPLAN 1999 conference on Programming language design and implementation, p.181-191, May 01-04, 1999, Atlanta, Georgia, United States
• 25: Polle T. Zellweger, An interactive high-level debugger for control-flow optimized programs (Summary), Proceedings of the ACM SIGSOFT/SIGPLAN software engineering symposium on High-level debugging, p.159-171, January 1983 [doi>10.1145/800007.808029]
• 26: Polle Trescott Zellweger, Interactive source-level debugging for optimized programs (compilation, high-level), University of California, Berkeley, 1984

INDEX TERMS

Primary Classification:
  D. Software
  D.2 SOFTWARE ENGINEERING
      D.2.5 Testing and Debugging
          Subjects: Debugging aids

General Terms:
Algorithms, Experimentation

Keywords:
compiler, debugger, non-transparent debugging, software-pipelining