Jingyan JOURDAN-LU

I am a PhD student in the Arénaire project-team, at LIP-ENS Lyon, under the supervision of Claude-Pierre Jeannerod, Christophe Monat and Jean-Michel Muller. My PhD thesis is under the convention of CIFRE, which aims to maximize the research collaboration between laboratory and industry. In my case, I also work as an engineer in the team, Compilation Expertise Center, at STMicroelectronics, Grenoble. email: lujingyan@gmail.com Curriculum-Vitae: , [PDF]

Research interests

• Fast floating-point computation on integer processors

  • Fused and specialized operators

  • Simultaneous sine and cosine

• Compilation assistance for more efficient floating-point applications

  • Static range analysis for FP specialization

  • Selection of fused operators

Publications

• Simultaneous floating-point sine and cosine for VLIW integer processors,
  with Claude-Pierre Jeannerod. Accepted for publication in the proceedings of ASAP 2012.

• Non-generic floating-point software support for embedded media processing,

with Claude-Pierre Jeannerod, Christophe Monat. Accepted for publication in the proceedings of SIES 2012.

• How to square floats accurately and efficiently on the ST231 integer processor,

  with Claude-Pierre Jeannerod, Christophe Monat, and Guillaume Revy. ARITH-20 Proceedings, Tübingen, July 2011.

• Techniques and tools for implementing IEEE 754 floating-point arithmetic on VLIW integer processors,
  with Christian Bertin, Claude-Pierre Jeannerod, Hervé Knochel, Christophe Monat, Christophe Mouilleron, Jean-Michel Muller, and Guillaume Revy. PASCO 2010 Proceedings, Grenoble, July 2010.

• NLMS-Based AEDA Algorithm for Sound Source Localization (Chinese),
  with Shou-zheng ZHU. Journal of East China Normal University, Natural Sc, 2007(5): 113-117.

Presentations

• 23rd IEEE International Conference on Applicationspecific Systems, Architectures and Processors,

  The Netherlands, July 2012 (planned).

• 7th IEEE International Symposium on Industrial Embedded Systems, Germany, June 2012 (planned).

• Arénaire Seminar, Ecole Normale Supérieure de Lyon, France, December 2011.

• DALI Seminar, Université de Perpignan, France, November 2011.

• 20th IEEE Symposium on Computer Arithmetic, Germany, July 2011. [slides]

Miscellaneous

• OpenCL: A standard framework to program your graphics card for general purpose computation.

Exhaustive testing for FP operators seems not possible for the moment. We are trying to use the massive computing power of GPU for our validation. Simply speaking, we want to test our operators on as many data as possible. Moreover, OpenCL specifies the built-in functions such as FMA and dot product, which can ease our testing.