A system configuration includes a processing element (PE), an input/output (I/O) interface device and a shared memory. The PE further includes at least one processing unit (PU) and one, or more, attached processing units (APUs). At least one of the APUs performs an I/O function by reading data from, and writing data to, an external device coupled to the I/O interface device. Data is exchanged between the APU and the I/O interface device via the shared memory using a data level synchronization mechanism.