GSRC Student Profile:
Daya Shanker Khudia
|
 |
(For more contact info, please log in.)
University of Michigan
Advisor: Valeria Bertacco
GSRC theme: resiliency
Expected graduation: Aug, 2014 |
Research Overview: Post-silicon debug and diagnosis automation for microprocessors
Due to increasing complexity and sheer size of multicore processors, many bugs escape pre-silicon verification either to be discovered during post-silicon validation or escaping to the field. Post-silicon validation engineers, mostly with limited design knowledge, are given the daunting task of
debugging with limited controllability and visibility of internal chip signals.
Moreover, many bugs during post-silicon validation don't manifest in every run of a testcase. Also the nondeterminism caused by varying memory access latency and other interrupts makes it very difficult to debug and diagnose bugs. My research focuses on developing solutions to automate the task of debugging and diagnosing during first silicon bring-up.
|
