The bottom line of any parallel system is the raw performance that can be delivered. Performance monitoring tools can improve understanding of a program's behaviour, particularly if the results can be visualised. Because each PE task executes an elaborated version of the sequential Haskell RTS, some of the profiling tools that are provided with sequential Haskell still yield information that can be used to tune parallel program. In addition, we have also produced special tools that give per-PE or per-thread activity profiles over time.