Feedback-Directed Barrier Optimization in a Strongly Isolated STM

摘要

Speed improvements in today's processors have largely been de-livered in the form of multiple cores, increasing the importance ofabstractions that ease parallel programming. Software transactionalmemory (STM) addresses many of the complications of concur-rency by providing a simple and composable model for safe accessto shared data structures. Software transactions extend a languagewith an atomic primitive that declares that the effects of a block ofcode should not be interleaved with actions executing concurrentlyon other threads. Adding barriers to shared memory accesses pro-vides atomicity, consistency and isolation. Strongly isolated STMs preserve the safety properties of trans-actions for all memory operations in a program, not just thoseinside an atomic block. Isolation barriers are added to non-transactional loads and stores in such a system to prevent thoseaccesses from observing or corrupting a partially completed trans-action. Strong isolation is especially important when integratingtransactions into an existing language and memory model. Isola-tion barriers have a prohibitive performance overhead, however, somost STM proposals have chosen not to provide strong isolation. In this paper we reduce the costs of strong isolation by cus-tomizing isolation barriers for their observed usage. The cus-tomized barriers provide accelerated execution by blocking threadswhose accesses do not follow the expected pattern. We use hotswap to tighten or loosen the hypothesized pattern, while pre-serving strong isolation. We introduce a family of optimizationhypotheses that balance verification cost against generality. We demonstrate the feasibility of dynamic barrier optimizationby implementing it in a bytecode-rewriting Java STM. Feedback-directed customization reduces the overhead of strong isolationfrom 505% to 38% across 11 non-transactional benchmarks; persis-tent feedback data further reduces the overhead to 16%. Dynamicoptimization accelerates a multi-threaded transactional benchmarkby 31% for weakly-isolated execution and 34% for strongly-isolated execution.