Unifying Thread-Level Speculation and Transactional Memory

Barreto, João; Dragojević, Aleksandar; Ferreira, Paulo J.; Filipe, Ricardo; Guerraoui, Rachid

doi:10.1007/978-3-642-35170-9_10

Cited by 18 publications

(17 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to the optimistic order, T2's read should return O T 1 a and T4's read should return O T 3 a . Even though this approach maximizes concurrency, its implementation requires traversing the shared lists of transactional meta-data, resulting in high transaction execution time and low performance [1].…”

Section: Parspecmentioning

confidence: 99%

Archie

Hirve

Palmieri

Ravindran

2014

Proceedings of the 15th International Middleware Conference on - Middleware '14

View full text Add to dashboard Cite

We present Archie, a high performance fault-tolerant transactional system. Archie complies with the State Machine Approach, where the transactional state is fully replicated and total ordered transactions are executed on the replicas. Archie avoids the serial execution after transactions get ordered, which is the typical bottleneck of those protocols, by anticipating the work and using speculation to process transactions in parallel, enforcing a predefined order. The key feature of Archie is to avoid any non-trivial operation to perform post total order's notification, in case the sequencer node remains stable (only a single timestamp increment is needed for committing a transaction). This approach significantly shortens the transaction's critical path. The contention of speculative execution is always kept limited by activating a fixed number of transactions at a time. A comprehensive evaluation, using three competitors and three well known benchmarks, shows that Archie outperforms competitors in all medium/high contention scenarios.

show abstract

Section: Parspecmentioning

confidence: 99%

Archie

Hirve

Palmieri

Ravindran

2014

Proceedings of the 15th International Middleware Conference on - Middleware '14

View full text Add to dashboard Cite

show abstract

“…Some of these works process transactions speculatively along di↵er-ent serialization orders [14,3], whereas others [16,1] fix a single order at the beginning and follow it. This intuition is the same as X-DUR but the deployment is di↵erent.…”

Section: Related Workmentioning

confidence: 99%

“…Speculation is a widely used technique for anticipating work based on uncertain inputs [3,14,16,16,1]. Some of these works process transactions speculatively along di↵er-ent serialization orders [14,3], whereas others [16,1] fix a single order at the beginning and follow it.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Speculative client execution in deferred update replication

Arun

Hirve

Palmieri

et al. 2014

Proceedings of the 9th Workshop on Middleware for Next Generation Internet Computing

View full text Add to dashboard Cite

Deferred Update Replication (DUR) is a powerful replication technique that allows parallelism of clients' execution while a global certification phase checks the validity of the transactional execution against workloads running on remote nodes. The well-known favorable scenario of DUR is when remote transactions rarely conflict with each other. In this paper we show that, even in this case, the conflicts happening among local application threads can significantly decrease performance. We address this problem by using speculation. We let local transactions propagate their postexecution snapshot to other local transactions before the outcome of the global certification is notified. This way, in scenarios where accesses are partitioned across nodes, we prevent local transactions from aborting each other. Through experimental study based on well-known transactional benchmarks we assess the e↵ectiveness of the approach, gaining more than 10⇥ using TPC-C benchmark.

show abstract

“…A remarkable property of TEPO is that a single thread can maintain several active transactions (thanks to decoupling execution from commit), in contrast to the previous proposals that support only one per thread (except TCC, which may support two transactions using double buffering). TLSTM [Barreto et al 2012] combines both TM and TLS approaches. All transactions execute unordered, but each one is split into several smaller tasks that are speculatively executed in parallel (ordered).…”

Section: Related Workmentioning

confidence: 99%

Effective Transactional Memory Execution Management for Improved Concurrency

Gonzalez-Mesa

Gutiérrez

Zapata

et al. 2014

ACM Trans. Archit. Code Optim.

View full text Add to dashboard Cite

This article describes a transactional memory execution model intended to exploit maximum parallelism from sequential and multithreaded programs. A program code section is partitioned into chunks that will be mapped onto threads and executed transactionally. These transactions run concurrently and out of order, trying to exploit maximum parallelism but managed by a specific fully distributed commit control to meet data dependencies. To accomplish correct parallel execution, a partial precedence order relation is derived from the program code section and/or defined by the programmer. When a conflict between chunks is eagerly detected, the precedence order relation is used to determine the best policy to solve the conflict that preserves the precedence order while maximizing concurrency. The model defines a new transactional state called executed but not committed. This state allows exploiting concurrency on two levels: intrathread and interthread. Intrathread concurrency is improved by having pending uncommitted transactions while executing a new one in the same thread. The new state improves interthread concurrency because it permits out-of-order transaction commits regarding the precedence order. Our model has been implemented in a lightweight software transactional memory system, TinySTM, and has been evaluated on a set of benchmarks obtaining an important performance improvement over the baseline TM system.

show abstract

Unifying Thread-Level Speculation and Transactional Memory

Cited by 18 publications

References 34 publications

Archie

Archie

Speculative client execution in deferred update replication

Effective Transactional Memory Execution Management for Improved Concurrency

Contact Info

Product

Resources

About