Memory tagging in Charm++

ABSTRACT

Many scientific applications are logically decomposed into modules. each module performing a different type of computation. These modules are then linked together inside the same executable. While these modules are logically independent, they are not physically independent: a faulty module can corrupt the state of another one.

By identifying the different modules inside an application, tagging the memory according to the different modules, and performing extra runtime checks, we can automatically detect certain type of errors. We implemented our idea inside the Charm++ runtime system, where modules can be easily identified. We illustrate the validity of our approach, and evaluate its overhead.

REFERENCES

• 1: A. Bhatele, S. Kumar, C. Mei, J. C. Phillips, G. Zheng, and L. V. Kale. Overcoming Scaling Challenges in Biomolecular Simulations across Multiple Platforms. In Proceedings of IEEE International Parallel and Distributed Processing Symposium 2008, 2008.
• 2: E. Bohm , A. Bhatele , L. V. Kalé , M. E. Tuckerman , S. Kumar , J. A. Gunnels , G. J. Martyna, Fine-grained parallelization of the Car-Parrinello ab initio molecular dynamics method on the IBM Blue Gene/L supercomputer, IBM Journal of Research and Development, v.52 n.1/2, p.159-175, January 2008
• 3: Intel Corporation. Intel Thread Checker. http://www.intel.com.
• 4: P. Jetley, F. Gioachin, C. Mendes, L. V. Kale, and T. R. Quinn. Massively Parallel Cosmological Simulations with ChaNGa. In Proceedings of IEEE International Parallel and Distributed Processing Symposium 2008, 2008.
• 5: R. Jyothi, O. S. Lawlor, and L. V. Kale. Debugging support for Charm++. In PADTAD Workshop for IPDPS 2004, page 294. IEEE Press, 2004.
• 6: L. V. Kalé. Performance and productivity in parallel programming via processor virtualization. In Proc. of the First Intl. Workshop on Productivity and Performance in High-End Computing (at HPCA 10), Madrid, Spain, February 2004.
• 7: L. V. Kale and S. Krishnan. Charm++: Parallel Programming with Message-Driven Objects. In G. V. Wilson and P. Lu, editors, Parallel Programming using C++, pages 175--213. MIT Press, 1996.
• 8: Sameer Kumar , Laxmikant V. Kale, Optimizing communication for massively parallel processing, University of Illinois at Urbana-Champaign, Champaign, IL, 2005
• 9: M. Ronsse and K. D. Bosschere. Non-intrusive on-the-fly data race detection using execution replay. In Automated and Algorithmic Debugging, 2000.
• 10: Paul Sack , Brian E. Bliss , Zhiqiang Ma , Paul Petersen , Josep Torrellas, Accurate and efficient filtering for the Intel thread checker race detector, Proceedings of the 1st workshop on Architectural and system support for improving software dependability, p.34-41, October 21-21, 2006, San Jose, California [doi>10.1145/1181309.1181315]
• 11: Julian Seward , Nicholas Nethercote, Using Valgrind to detect undefined value errors with bit-precision, Proceedings of the annual conference on USENIX Annual Technical Conference, p.2-2, April 10-15, 2005, Anaheim, CA
• 12: T. Technologies. Debugging Memory Problems Using TotalView Debugger. http://www.totalviewtech.com.
• 13: T. Technologies. TotalView® debugger. http://www.totalviewtech.com/TotalView.
• 14: Gengbin Zheng , Laxmikant V. Kale, Achieving high performance on extremely large parallel machines: performance prediction and load balancing, University of Illinois at Urbana-Champaign, Champaign, IL, 2005

INDEX TERMS

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

Additional Classification:
  D. Software
  D.2 SOFTWARE ENGINEERING
      D.2.2 Design Tools and Techniques
          Subjects: Modules and interfaces

General Terms:
Design, Experimentation, Measurement, Performance

Keywords:
debugging, memory debugging, memory tagging, parallel debugging