Location

UNLV SEB Lobby & Auditorium

Start Date

26-4-2013 9:30 AM

End Date

26-4-2013 3:00 PM

Description

As CMOS technology scales down, process variation introduces significant uncertainty in power and performance to VLSI circuits and significantly affects their reliability. Although Static-Timing Analysis (STA) remains an excellent tool, current trends in process scaling have imposed significant difficulties to STA. As one of the promising solutions, Statistical static timing analysis (SSTA) has become the frontier research topic in recent years in combating such variation effects. This poster will be focusing on two aspects of SSTA and its applications in VLSI designs: (1) Statistical timing modeling and analysis; and (2) Architectural implementations of the atomic operations (max and add) using Vector Thread programming Model. Experimental results have shown that our approach can provide 282 times speedup when compared to a conventional CPU implementation.

Keywords

Computer architecture; Electronic data processing; Metal oxide semiconductors; Complementary; Threads (Computer programs)

Disciplines

Computer Engineering | Computer Sciences | Signal Processing | Systems Architecture | VLSI and Circuits, Embedded and Hardware Systems

Language

English


Share

COinS
 
Apr 26th, 9:30 AM Apr 26th, 3:00 PM

Modeling and Architectural Simulations of the Statistical Static Timing Analysis of the Variation Sources For VLSI Circuits

UNLV SEB Lobby & Auditorium

As CMOS technology scales down, process variation introduces significant uncertainty in power and performance to VLSI circuits and significantly affects their reliability. Although Static-Timing Analysis (STA) remains an excellent tool, current trends in process scaling have imposed significant difficulties to STA. As one of the promising solutions, Statistical static timing analysis (SSTA) has become the frontier research topic in recent years in combating such variation effects. This poster will be focusing on two aspects of SSTA and its applications in VLSI designs: (1) Statistical timing modeling and analysis; and (2) Architectural implementations of the atomic operations (max and add) using Vector Thread programming Model. Experimental results have shown that our approach can provide 282 times speedup when compared to a conventional CPU implementation.