Citation
Aaron Ng, Igor Markov. "Toward Quality Tools and Tool Flows through High-performance Computing". Sixth International Symposium on Quality of Electronic Design (ISQED'05), pp. 22-27, March, 2005.

Abstract
As the scale and complexity of VLSI circuits increase, Electronic Design Automation (EDA) tools become much more sophisticated and are held to increasing standards of quality. New-generation EDA tools must work correctly on a wider range of inputs, have more internal states, take more effort to develop, and offer fertile ground for programming mistakes. Ensuring quality of a commercial tool in realistic design flows requires rigorous simulation, non-trivial computational resources, accurate reporting of results and insightful analysis. However, time-to-market pressures encourage EDA engineers and chip designers to look elsewhere. Thus, the recent availability of cheap Linux clusters and grids shifts the bottleneck from hardware to logistical tasks, i.e., the speedy collection, reporting and analysis of empirical results. To be practically feasible, such tasks must be automated; they leverage high-performance computing to improve EDA tools. In this work we outline a possible infrastructure solution, called bX, explore relevant use models and describe our computational experience. In a specific application, we use bX to automatically build Pareto curves required for accurate performance analysis of randomized algorithms.

Electronic downloads

• isqed05-bx.pdf · application/pdf · 122 kbytes

Citation formats

• HTML

  Aaron Ng, Igor Markov. <a
  href="http://www.gigascale.org/pubs/599.html">Toward
  Quality Tools and Tool Flows through High-performance
  Computing</a>, Sixth International Symposium on
  Quality of Electronic Design (ISQED'05), pp. 22-27, March,
  2005.

• Plain text

  Aaron Ng, Igor Markov. "Toward Quality Tools and Tool Flows
  through High-performance Computing". Sixth International
  Symposium on Quality of Electronic Design (ISQED'05), pp.
  22-27, March, 2005.

• BibTeX

  @inproceedings{NgMarkov05_TowardQualityToolsToolFlowsThroughHighperformanceComputing,
      author = {Aaron Ng and Igor Markov},
      title = {Toward Quality Tools and Tool Flows through
                High-performance Computing},
      booktitle = {Sixth International Symposium on Quality of
                Electronic Design (ISQED'05)},
      pages = {pp. 22-27},
      month = {March},
      year = {2005},
      abstract = {As the scale and complexity of VLSI circuits
                increase, Electronic Design Automation (EDA) tools
                become much more sophisticated and are held to
                increasing standards of quality. New-generation
                EDA tools must work correctly on a wider range of
                inputs, have more internal states, take more
                effort to develop, and offer fertile ground for
                programming mistakes. Ensuring quality of a
                commercial tool in realistic design flows requires
                rigorous simulation, non-trivial computational
                resources, accurate reporting of results and
                insightful analysis. However, time-to-market
                pressures encourage EDA engineers and chip
                designers to look elsewhere. Thus, the recent
                availability of cheap Linux clusters and grids
                shifts the bottleneck from hardware to logistical
                tasks, i.e., the speedy collection, reporting and
                analysis of empirical results. To be practically
                feasible, such tasks must be automated; they
                leverage high-performance computing to improve EDA
                tools. In this work we outline a possible
                infrastructure solution, called bX, explore
                relevant use models and describe our computational
                experience. In a specific application, we use bX
                to automatically build Pareto curves required for
                accurate performance analysis of randomized
                algorithms.},
      URL = {http://www.gigascale.org/pubs/599.html}
  }
  

Posted by Aaron Ng on 29 Mar 2005..

Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.