Local (V,T) estimation in integrated circuits using a set of statistical tests

Người trình bày: Giáo sư Suzanne Lesecq, CEA-LETI, MINATEC, Cộng hoà Pháp

Thời gian: Từ 13h30 đến 14h30, ngày 12/10/2012

Địa điểm: Phòng 212, nhà E3, Trường Đại học Công nghệ – 144 đường Xuân Thuỷ, quận Cậu Giấy, Hà Nội.

Tóm tắt:

Mobile platforms need ever-increasing computational performances under stringent energy consumption limitation mainly due to the battery lifespan.  An  optimal  operating  point  is  obtained  thanks  to  a  compromise  between  performance  and  power consumption. For distributed architectures (e.g. MultiProcessor System on Chip), the supply voltage  and  the  operating  frequency  of  each  processing  element  are  dynamically  tuned  to reach efficient performance/power consumption trade-offs. To increase the performance of each “actuator”, the physical state (e.g.  the  current  supply  voltage  and  temperature)  of  the integrated  circuit  must  be  monitored  to  locally  adapt  the  control  parameters.  During this keynote, we will present a new estimation method based on statistical tests to estimate the supply  voltage  and  the  temperature  of  a  local  area  in  an  integrated  circuit.  Standard ring  oscillators  buried  in  the  chip provide the raw measurements that are fused to estimate the IC physical state. The  results  presented  here  have  been  developed  in  cooperation  between  the  CEA-LETI (Mrs E. Beigné, Mr. L. Vincent, Dr. S. Lesecq), Grenoble, France and the LIRMM, Montpellier, France (Dr. Ph. Maurines).

Tiểu sử

Suzanne Lesecq defended a PhD in Process Control from the Grenoble Institute of Technology, France, in 1997. She has been appointed as Associate-Professor from 1998 to 2006 and full-time Professor from 2006 to 2009 at the University of Grenoble. There she was head of the Research team SA-IGA at the GIPSA-Lab (Grenoble Image – Signal Processing – Automatic  control Lab), and conducted research dedicated to Fault Detection and Isolation (FDI). She joined CEA-LETI in mid-2009 where she is conducting research on “Automatic control for computing”, more precisely, related to control theory applied to MPSoC with application to power management, and motion capture based on data fusion techniques.