Abstract
dc:descriptionComputing platforms increasingly incorporate heterogeneous memory hardware technologies, as a way to scale application performance, memory capacities and achieve cost-effectiveness. However, this heterogeneity, along with the greater irregularity in the behavior of emerging workloads, render existing hybrid memory management approaches ineffective, calling for more intelligent methods. To this end, this thesis reveals new insights, develops novel methods and contributes system-level mechanisms towards the practical integration of machine learning into hybrid memory management, boosting application performance and system resource efficiency. The specific contributions of this thesis are as follows.First, this thesis builds Kleio, a hybrid memory page scheduler with machine intelligence. Kleio deploys Recurrent Neural Networks to learn memory access patterns at a page granularity and improve upon the selection of dynamic page migrations across the memory hardware components. Kleio cleverly focuses the machine learning on the page subset whose timely movement will reveal most application performance improvement, while preserving history-based lightweight management for the rest of the pages. In this way, Kleio bridges on average 80% of the relative existing performance gap, while laying the grounds for practical machine intelligent data management with manageable learning overheads.Second, this thesis contributes Cori, a system-level solution for tuning the operational frequency of periodic page schedulers for hybrid memories. Cori synthesizes information on data reuse to properly identify the data movement frequencies to be tested, reducing by 5× the number of tuning trials compared to existing empirical or insight-less tuning approaches. In this way, Cori delivers application performance levels within 3% from the case of optimally selected frequency, eliminating the 10%-100% performance gap created when using frequencies currently adopted by related works. Such improvements are complimentary to the use of machine learning and further boost its effect on application performance. Third, this thesis contributes Coeus, a page grouping mechanism for hybrid memory page schedulers with machine intelligence, such as Kleio. Coeus leverages the data reuse insights revealed by Cori to fine-tune the granularity at which patterns are interpreted by the page scheduler, increasing the pattern similarity across pages. Then, Coeus groups together the pages that share the same access behavior, enabling the training of a single Recurrent Neural Network per page cluster. As a result, Coeus reduces by almost 3 times the associated learning overheads compared to Kleio. In addition, Coeus achieves 3 times higher application performance, by the combined effects of applying machine learning to more pages and by performing management operations at a fine-tuned granularity.Finally, this thesis contributes Cronus, an image-based page selector for hybrid memory page schedulers with machine intelligence, such as Kleio. Cronus uses visualization to accelerate the process of selecting which page patterns should be managed with machine learning. The visualization reveals spatial and temporal correlations across pages, that Kleio cannot capture during its page selection process. Instead, Kleio uses elaborate performance estimation models that come with non-trivial operational overheads. Cronus builds a lightweight visualization pipeline that detects page access patterns for machine learning-based management. The quality of the selected pages is comparable to Kleio’s and delivers similar levels of application performance, in return for 75 times reduction in the selection time. Cronus lays the foundations for future use of visualization and computer vision methods in memory management, such as image-based memory access pattern classification, recognition and prediction.
Degree
thesis:*- Grantor dc:publisher
- Institutes outside Greece
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Doudali, Thaleia-Dimitra
- Δούδαλη, Θάλεια-Δήμητρα
Subjects
dc:subject × 22- Διαχείριση μνήμης
- Υβριδική μνήμη
- Μη πτητικές μνήμες
- Τεχνικές διαχείρισης πόρων
- Μηχανική μάθηση
- Τεχνητή νοημοσύνη
- Επαναλαμβανόμενα νευρωνικά δίκτυα
- Ευριστικές μέθοδοι
- Memory management
- Hybrid memory
- Non volatile memories
- Resource management techniques
- Machine learning
- Artificial intelligence
- Recurrent neural networks
- Heuristic algorithms
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Υπολογιστές, Υλικό (hardware) και Αρχιτεκτονική
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Computer science, Hardware and Architecture
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/53218