Multicores are now the norm and new generations of software must take advantage of the presence of several cores in a given architecture. Parallel programming requires specific skills beyond from those required for the development of traditional sequential programs. The usage of parallel libraries is one of the best ways to facilitate parallel programming, as they does not require new compilers and they allow to parallelize sequential codes without big efforts by the programmer. This paper presents an efficient and portable parallel set container. This container has been used to program parallel versions of several algorithms. The experimental results show that the container facilitates the programmability and achieves a good performance on multicore systems.