Return to: U of M Home

The SEASONing Tool: A Spice Engine for Adding Soft-errors On Netlists

The SEASONing tool was created by AJ KleinOsowski, Kevin KleinOsowski, Mahesh Subramony, and David J. Lilja.

Direct comments, questions, and clarifications to Professor David J. Lilja at lilja at umn.edu.

In this project, we describe a technique and the associated software tool for injecting and modeling transient radiation events, so-called soft-errors, in Spice netlists. We create a parser written in C to read and analyze a Spice netlist text file. The parser chooses random nodes and adds current pulses to these nodes to mimic the charge generated from a high energy particle strike. The fault density and time windows of the faults are fully configurable. The resulting modified netlist is then simulated using Spice. With our netlist parser, called SEASONing, designers can evaluate the robustness of their circuits. This produces a new metric by which alternative designs may be evaluated for soft-error robustness.

Please reference this technical report if you use any of this work:
The SEASONing Tool: A Spice Engine for Adding Soft-errors On Netlists , Kevin KleinOsowski, AJ KleinOsowski, and David J. Lilja, University of Minnesota Laboratory for Advanced Research in Computing Technology and Compilers Technical Report No. ARCTiC 05-09, May, 2005.

The code may be downloaded from:SEASONing.

---

This work was supported in part by IBM, Semiconductor Research Corporation contract no. 2004-HJ-1190, National Science Foundation grant CCR-0210197, the University of Minnesota Digital Technology Center and the Minnesota Supercomputing Institute.