Seer

Abstract

Scheduling concurrent transactions to minimize contention is a well known technique in the Transactional Memory (TM) literature, which was largely investigated in the context of software TMs. However, the recent advent of Hardware Transactional Memory (HTM), and its inherently restricted nature, pose new technical challenges that prevent the adoption of existing schedulers: unlike software implementations of TM, existing HTMs provide no information on which data item or contending transaction caused abort.We propose Seer, a scheduler that addresses precisely this restriction of HTM by leveraging on an on-line probabilistic inference technique that identifies the most likely conflict relations, and establishes a dynamic locking scheme to serialize transactions in a fine-grained manner. Our evaluation shows that Seer improves the performance of the Intel TSX HTM by up to 2.5×, and by 62% on average, in TM benchmarks with 8 threads. These performance gains are not only a consequence of the reduced aborts, but also of the reduced activation of the HTM's pessimistic fall-back path.

Migration in Hardware Transactional Memory on Asymmetric Multiprocessor

“…This solution assumes that transaction code is built from an already existing lock-based code and programmer-added locks are used as auxiliary locks. In [15] when transaction commits or aborts, data are collected which other transactions are active. If it is a commit that means that active transactions do not conflict and can be executed in parallel, otherwise they should not be scheduled at the same time.…”