Abstract:
In order to detect and locate defects, or faults, in a plurality of chips or other circuits sharing a common design, said chips are each tested for incorrect outputs, or failures, in response to inputs. The incorrect outputs are then collectively diagnosed in a single simulation by simulating a series of suspected fault candidates on a simulated chip of the chip design, and afterward comparing the incorrect outputs generated by each fault candidate to the incorrect outputs of the individual chips, to determine if a fault candidate generates all failures for a chip and no others. The test inputs and expected outputs may be predetermined through Automatic Test Pattern Generation. The fault candidates may be determined by use of a backtrace process such as back cone tracing. The failures may be recorded in association with a measure point, the input pattern that resulted in the failure, and the failure value.