Multiuser diversity is a form of diversity that is inherent in wireless systems with more than one user. For single-best user selection, it becomes a scheme whereby the base station schedules transmission to and from the user with the best channel. Traditionally, this form of diversity is considered in conjunction with adaptive modulation, where the rate of transmission is adapted to the channel quality of the best user. In this paper, we consider a fixed modulation and analyze the asymptotic bit-error-rate performance for uplink systems for a large number of users using the theory of asymptotic order statistics. Moreover, we consider multiple antennas at the base station and show that antenna correlation, perhaps surprisingly, improves the bit error rate (BER) for a large number of users. Simulation results corroborate our analytical results.