HEUSPEC: A Software Speculation Parallel Model

Xu, Fan; Shen, Li; Wang, Zhiying; Guo, Hui; Su, Bo; Chen, Wei

doi:10.1109/icpp.2013.76

Cited by 5 publications

(6 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We implemented the three optimizations in the runtime library of HEUSPEC [3] , a software speculative parallel model. Fig.11 shows the modified part in the original layers of the runtime library.…”

Section: Methodsmentioning

confidence: 99%

“…Tian et al [1] applied pre-computing value prediction (PCVP) to make the prediction more accurate, with additional learning process and compiling supports. Xu et al [3] applied heuristic value prediction (HVP) to make a compromise between the former two optimizations, predicting the values of conflict variables with their history value traces. Ding et al [2] proposed BoP with value based correctness checking technique, to avoid some unnecessary rollbacks of speculative tasks.…”

Section: Related Workmentioning

confidence: 99%

“…For example, the value prediction technique [4] and the value-based correctness checking technique [2] could be applied to reduce miss rate. The incremental recovery technique [5] could be applied to reduce O rbk , and the granularity resizing technique [3] could be applied to reduce O ctrl .…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, to reduce the global overhead has become the key point to improve the performance of software speculative parallel models. The global overhead of a software speculative parallel model is tightly related to three factors [3] : the average control overhead, the rollback overhead, and the mis-speculation rate. For given number of speculative tasks, the global overhead can be represented as follows:…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Optimization Strategies Oriented to Loop Characteristics in Software Thread Level Speculation Systems

Shen

Wang

2016

J. Comput. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

Thread level speculation provides not only a simple parallel programming model, but also an effective mechanism for thread-level parallelism exploitation. The performance of software speculative parallel models is limited by high global overheads caused by different types of loops. These loops usually have different characteristics of dependencies and different requirements of optimization strategies. In this paper, we propose three comprehensive optimization techniques to reduce different factors of global overheads, aiming at requirements from different types of loops. Inter-thread fetching can reduce the high mis-speculation rate of the loops with frequent dependencies and out-of-order committing can reduce the control overhead of the loops with infrequent dependencies, while enhanced dynamic task granularity resizing can reduce the control overhead and optimize the global overhead of the loops with changing characteristics of dependencies. All these three optimization techniques have been implemented in HEUSPEC, a software TLS system. Experimental results indicate that they can satisfy the demands from different groups of benchmarks. The combination of these techniques can improve the performance of all benchmarks and reach a higher average speedup.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Optimization Strategies Oriented to Loop Characteristics in Software Thread Level Speculation Systems

Shen

Wang

2016

J. Comput. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…HEUSPEC [14] is a software speculation parallel model that dynamically adapts to different value predictors and granularity tasks. While adaptive forkheuristics target the problem of fork point selection.…”

Section: Related Workmentioning

confidence: 99%

Adaptive Fork-Heuristics for Software Thread-Level Speculation

Cao

Verbrugge

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Fork-heuristics play a key role in software Thread-Level Speculation (TLS). Current fork-heuristics either lack real parallel execution environment information to accurately evaluate fork points and/or focus on hardware-TLS implementation which cannot be directly applied to software TLS. This paper proposes adaptive fork-heuristics as well as a feedback-based selection technique to overcome the problems. Adaptive fork-heuristics insert and speculate on all potential fork/join points and purely rely on the runtime system to disable inappropriate ones. Feedback-based selection produces parallelized programs with ideal speedups using log files generated by adaptive heuristics. Experiments of three scientific computing benchmarks on a 64-core machine show that feedback-based selection and adaptive heuristics achieve more than 88% and 50% speedups of the manual-parallel version, respectively. For the Barnes-Hut benchmark, feedback-based selection is 49% faster than the manual-parallel version.

show abstract

Improving Speculation Accuracy with Inter-thread Fetching Value Prediction

Shen

Wang

et al. 2014

Algorithms and Architectures for Parallel Processing

View full text Add to dashboard Cite

HEUSPEC: A Software Speculation Parallel Model

Cited by 5 publications

References 17 publications

Optimization Strategies Oriented to Loop Characteristics in Software Thread Level Speculation Systems

Optimization Strategies Oriented to Loop Characteristics in Software Thread Level Speculation Systems

Adaptive Fork-Heuristics for Software Thread-Level Speculation

Improving Speculation Accuracy with Inter-thread Fetching Value Prediction

Contact Info

Product

Resources

About