Fault Simulation Using Irsim

In order to run a fault simulation, you must use a special version of Irsim, called ``ifsim''. If you previously used Irsim to simulate your design, you only need to create one additional file before starting a fault simulation. Ifsim accepts all the same commands as Irsim, plus one additional command to initiate a fault simulation: faultsim.

The easiest way to run a fault simulation on your design is to first run Ifsim as you would using Irsim, and at the end of the simulation run issue the faultsim command. This command has the following form:


		faultsim setup-file 
$\left[\vphantom{\mbox{ \em output-file }}\right.$ output-file $\left.\vphantom{\mbox{ \em output-file }}\right]$

where setup-file is the name of a file containing the information regarding output pins and their sampling timing, and output-file is the name of the file where the fault-simulation output will be recorded. If no output-file is specified, Ifsim will use the file fsim.out.

After reading the setup file, Ifsim will select the maximum number of nodes into which stuck-at faults will be injected. By default, Ifsim uses 20% of all nodes in the circuit. Fault seeding then proceeds by randomly selecting the required number of nodes from all candidate nodes. A node is a fault candidate if at least one transistor gate is connected to the node, and the node is not a primary input (it is never driven). At the end of the simulation, the output file will contain detailed information regarding each fault tested; this is followed by a brief summary that includes the fault coverage.



Subsections