Software Hint-Driven Data Management for Hybrid Memory in Mobile Systems

Abstract

Hybrid memory systems, comprised of emerging non-volatile memory (NVM) and DRAM, have been proposed to address the growing memory demand of current mobile applications. Recently emerging NVM technologies, such as phase-change memories (PCM), memristor, and 3D XPoint, have higher capacity density, minimal static power consumption and lower cost per GB. However, NVM has longer access latency and limited write endurance as opposed to DRAM. The different characteristics of distinct memory classes render a new challenge for memory system design. Ideally, pages should be placed or migrated between the two types of memories according to the data objects’ access properties. Prior system software approaches exploit the program information from OS but at the cost of high software latency incurred by related kernel processes. Hardware approaches can avoid these latencies, however, hardware’s vision is constrained to a short time window of recent memory requests, due to the limited on-chip resources. In this work, we propose OpenMem: a hardware-software cooperative approach that combines the execution time advantages of pure hardware approaches with the data object properties in a global scope. First, we built a hardware-based memory manager unit (HMMU) that can learn the short-term access patterns by online profiling, and execute data migration efficiently. Then, we built a heap memory manager for the heterogeneous memory systems that allows the programmer to directly customize each data object’s allocation to a favorable memory device within the presumed object life cycle. With the programmer’s hints guiding the data placement at allocation time, data objects with similar properties will be congregated to reduce unnecessary page migrations. We implemented the whole system on the FPGA board with embedded ARM processors. In testing under a set of benchmark applications from SPEC 2017 and PARSEC, experimental results show that OpenMem reduces 44.6% energy consumption with only a 16% performance degradation compared to the all-DRAM memory system. The amount of writes to the NVM is reduced by 14% versus the HMMU-only, extending the NVM device lifetime.

A Computer Vision-Based Automatic System for Egg Grading and Defect Detection

“…Despite the research;s high performance in sorting egg quality based on egg surface and weight, some further studies could the model be applied to real-world situations: (a) using emerging nonvolatile memory (NVM) to reduce memory footprint and latency [49], which is crucial for mobile application; (b) extending the model to egg datasets with more diversity (other egg colors, egg multiplication and other spices) to fulfill the application environment; (c) using a 360-degree camera to prevent misidentification in cracked and bloody eggs; (d) optimize the sorting and weighing process to reduce the time required to complete the task without sacrificing accuracy; (e) enhancing the accuracy of egg segmentation by leveraging the segment-anything model [50].…”

Semi-supervised Graph Instance Transformer for Mental Health Inference

“…Short term monitoring cannot guarantee capturing participants' symptoms and their associated behavior changes when they are unwell. Meanwhile, if we retrain models using all existing data, the computation and memory burden would be high and prohibitive for mobile devices [11][12][13]. If we retrain models using only the newly incoming data, a classic problem referred to as catastrophic forgetting will significantly downgrade model performance.…”

“…Incremental learning is designed for models built on streaming data to mitigate the catastrophic forgetting problem [14] by continuously updating the models with knowledge transfer. It also helps the model continuously monitor users' health states, while adapt to computation and memory constraints in mobile devices [13,15].…”