Learning-based Phase-aware Multi-core CPU Workload Forecasting

Abstract

Predicting workload behavior during workload execution is essential for dynamic resource optimization in multi-processor systems. Recent studies have proposed advanced machine learning techniques for dynamic workload prediction. Workload prediction can be cast as a time series forecasting problem. However, traditional forecasting models struggle to predict abrupt workload changes. These changes occur because workloads are known to go through phases. Prior work has investigated machine learning-based approaches for phase detection and prediction, but such approaches have not been studied in the context of dynamic workload forecasting. In this paper, we propose phase-aware CPU workload forecasting as a novel approach that applies long-term phase prediction to improve the accuracy of short-term workload forecasting. Phase-aware forecasting requires machine learning models for phase classification, phase prediction, and phase-based forecasting that have not been explored in this combination before. Furthermore, existing prediction approaches have only been studied in single-core settings. This work explores phase-aware workload forecasting with multi-threaded workloads running on multi-core systems. We propose different multi-core settings differentiated by the number of cores they access and whether they produce specialized or global outputs per core. We study various advanced machine learning models for phase classification, phase prediction, and phase-based forecasting in isolation and different combinations for each setting. We apply our approach to forecasting of multi-threaded Parsec and SPEC workloads running on an 8-core Intel Core-i9 platform. Our results show that combining GMM clustering with LSTMs for phase prediction and phase-based forecasting yields the best phase-aware forecasting results. An approach that uses specialized models per core achieves an average error of 23% with up to 22% improvement in prediction accuracy compared to a phase-unaware setup.

Evolutionary Game Theory-Based Optimal Scheduling Strategy for Heterogeneous Computing

“…In [29], a deep Q-learning resource prediction and scheduling algorithm for GPU is proposed, which designed three prototypes of resource management systems, the simulation results show significant improvements in resource utilization compared to ordinary heuristics. The authors in [30] proposed a deep learning based on multi-core CPU workload prediction by fusing GMM clustering with LSTM algorithm for phase prediction, which will produce the best phase-aware prediction results and reduces the average error.…”

“…The heterogeneous computing resource scheduling has been studied in [22], [23], [24], and [25], but it mainly considers the static resource scheduling scenario, the temporary resource switching scenario is ignored. Meanwhile, although resource prediction has effectively improved the flexible scheduling ability of computing resources in [26], [27], [28], [29], and [30], but not consider actual costs and profit from the perspective of operators and how to maximize the benefits of computing operations. Besides, the current research on resource scheduling based on game theory is mainly oriented to cloud computing power pricing, network elements, and other fields [31], [32], [33], and [34].…”

“…From formula (30), it can be seen that when the rental ratio of the CPTP is 0, the CPTP will not rent computing services. HCSP and CPTP will not benefit from the computing power network, and the quality of computing power application providers cannot be improved.…”