Abstract: A method, apparatus and system for integrated circuit testing, wherein a plural number of devices under test (DUTs) and a plural number of comparison apparatuses are placed on a common substrate. The DUTs all operate under the same input stimulation and each produce its own operation output. The outputs are compared by the comparison apparatuses to generate comparison characteristics which are used to filter-out the failed devices. This invention lowers the testing cost, shortens time to product mass-production, and lowers the miss rate of failed devices passed as good ones.