Berkeley FIT estimation tool (BFIT)
BFIT is a tool that can determine the neutron FIT (failure-in-time, per 109 hours) contribution of each combinational gate in a sequential circuit. It is created by Dan Holcomb, Wenchao Li, and Professor Sanjit Seshia as part of the Verification-Guided Error Resilience (VGER) project at UC Berkeley.
Combinational gates cause FIT when a neutron strike happens in the vicinity of a reverse-biased diffusion junction, causing a spurious voltage glitch which propagates to a downstream sequential where it can be latched into system state.
BFIT works by applying an input vector to the circuit and then determining the FIT contribution of each gate when that vector is applied. This process is repeated for many vectors to determine the overall FIT contribution of each gate. Thousands of vectors can be analyzed in just minutes on circuits with up to 20,000 gates.
D. Holcomb, W. Li, S. A. Seshia, Design as you see FIT: System-Level Soft Error Analysis of Sequential Circuits, Design Automation and Test in Europe, 2009 (to appear) [paper]
The development of BFIT is supported in part by the Gigascale Systems Research Center (GSRC).