The problem of self-diagnosis of multiprocessor and multicomputer systems under the generalized comparison model (GCM) is considered. GCM assumes that a set of jobs is assigned to pairs of units and that the outcomes are compared by the units themselves (selfdiagnosis). Based on the set of comparison outcomes (agreements and disagreements among the units), the set of up to t faulty nodes is identified (t-diagnosable systems). This paper proposes an artificial-immunebased algorithm to solve the fault identification problem. The immune diagnosis algorithm correctly identifies the set of faulty units, and it has been evaluated using randomly generated t-diagnosable systems. Simulation results indicate that the proposed approach is a viable alternative to solve the GCM-based diagnosis problem.