Multithreaded processors, in the context of real-time sys- tems, create the opportunity to mix, on the same hardware platform, the execution of a complex real-time workload and the execution of non-critical applications. But resources sharing, inherent to multithreading, hinders the timing anal- ysis of concurrent tasks. Such analyses are critical to real- time tasks which have timing deadlines that must be met and enforced. In this paper, we present the PRETI, Partitioned REal- TIme shared cache scheme, a flexible, low implementation- overhead, shared cache partitioning scheme. PRETI can preclude inter-task conflicts on shared caches, and their pes- simistic impact on timing estimates, by allocating private cache space to real-time tasks. Therefore, uniprocessor, i.e. unithread, worst-case execution time (WCET) estimation techniques can be used. The remaining cache space, not re- served for currently running real-time tasks, is shared by all tasks running on the processor, in particular the non-critical ones, enabling high performances for these tasks. Experiments are presented to show that the PRETI cache scheme allows for easing and guaranteeing the schedulability of a set of real-time tasks with tight timing constraints, and enabling high performance for the non-critical tasks.