Discovering And Understanding Performance Bottlenecks In Transactional Applications

Many researchers have developed applications using transactional memory (TM) with the purpose of benchmarking different implementations, and studying whether or not TM is easy to use. However, comparatively little has been done to provide general-purpose tools for profiling and tuning programs which use transactions.In this paper we introduce a series of profiling techniques for TM applications that provide in-depth and comprehensive information about the wasted work caused by aborting transactions. We explore three directions: (i) techniques to identify multiple potential conflicts from a single program run, (ii) techniques to identify the data structures involved in conflicts by using a symbolic path through the heap, rather than a machine address, and (iii) visualization techniques to summarize how threads spend their time and which of their transactions conflict most frequently.To examine the effectiveness of the profiling techniques, we provide a series of illustrations from the STAMP TM benchmark suite and from the synthetic Worm Bench workload. We show how to use our profiling techniques to optimize the performance of the Bayes, Labyrinth and Intruder applications.We discuss the design and implementation of our techniques in the Bartok-STM system. We process data offline or during garbage collection, where possible, in order to minimize the probe effect introduced by profiling.