Machine learning-based thread-parallelism regulation in software transactional memory

Rughetti, Diego; Sanzo, Pierangelo Di; Ciciani, Bruno; Quaglia, Francesco

doi:10.1016/j.jpdc.2017.06.001

Cited by 3 publications

(4 citation statements)

References 38 publications

(33 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The analytical model used by the thread scheduling technique in [21] is instantiated via regression analysis applied to a family of reference functions, based on measurements preventively collected by profiling the workload. Similarly, the neural network model used in [2] requires a wide training set built via a workload profiling phase. This technique was improved in [22] through a dynamic feature selection mechanism.…”

Section: Related Workmentioning

confidence: 99%

“…(see Figure 1) that we measured while running the abovementioned benchmark applications 2 . The plot shows that the energy consumption curves are quite similar to the ones related to the execution time.…”

Section: Preliminary Experimental Analysismentioning

confidence: 99%

“…Thread scheduling techniques (e.g. [2]) are instead based on controlling the number of threads used to run the application. The objective is therefore the one of regulating the thread-level parallelism so as to maximize the transaction throughput.…”

Section: Introductionmentioning

confidence: 99%

“…In any case, given that the workload profile of the applications can be unknown and/or may change along the application lifetime, scheduling techniques For what concerns Software TM (STM) systems, literature scheduling techniques [3] can be separated into two groups: (a) the ones based on performance prediction models (e.g. [2]) and (b) the ones based on heuristic approaches (e.g. [4]).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Adaptive Model-Based Scheduling in Software Transactional Memory

Sanzo

Pellegrini

Sannicandro

et al. 2020

IEEE Trans. Comput.

Self Cite

View full text Add to dashboard Cite

Software Transactional Memory (STM) stands as powerful concurrent programming paradigm, enabling atomicity and isolation while accessing shared data. On the downside, STM may suffer from performance degradation due to excessive conflicts among concurrent transactions, which cause waste of CPU-cycles and energy because of transaction aborts. An approach to cope with this issue consists of putting in place smart scheduling strategies which temporarily suspend the execution of some transaction in order to reduce the transaction conflict rate. In this article, we present an adaptive model-based transaction scheduling technique relying on a Markov Chain-based performance model of STM systems. Our scheduling technique is adaptive in a twofold sense: (i) it controls the execution of transactions depending on throughput predictions by the model as a function of the current system state, (ii) it re-tunes on-line the Markov Chain-based model to adapt it-and the outcoming transaction scheduling decisions-to dynamic variations of the workload. We have been able to achieve the latter target thanks to the fact that our performance model is extremely lightweight. In fact, to be recomputed, it requires a reduced set of input parameters, whose values can be estimated via a few on-line samples related to the current workload dynamics. We also present a scheduler that implements our adaptive technique, which we integrated within the open source TinySTM package. Further, we report the results of an experimental study based on the STAMP benchmark suite, which has been aimed at assessing both the accuracy of our performance model in predicting the actual system throughput and the advantages of the adaptive scheduling policy over literature techniques.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Preliminary Experimental Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations