scholarly journals Power-Time Exploration Tools for NMP-Enabled Systems

Electronics ◽  
2019 ◽  
Vol 8 (10) ◽  
pp. 1096
Author(s):  
Chae Eun Rhee ◽  
Seung-Won Park ◽  
Jungwoo Choi ◽  
Hyunmin Jung ◽  
Hyuk-Jae Lee
Keyword(s):  
Memory Performance ◽  
Main Memory ◽  
Design Decision ◽  
Evaluation Tool ◽  
Memory Processing ◽  
Hybrid Memory ◽  
System A ◽  
Trade Offs ◽  
High Bandwidth ◽  
Effective Use

Recently, dramatic improvements in memory performance have been highly required for data demanding application services such as deep learning, big data, and immersive videos. To this end, the throughput-oriented memory such as high bandwidth memory (HBM) and hybrid memory cube (HMC) has been introduced to provide a high bandwidth. For its effective use, various research efforts have been conducted. Among them, the near-memory-processing (NMP) is a concept that utilizes bandwidth and power consumption by placing computation logic near the memory. In the NMP-enabled system, a processor hierarchy consisting of hosts and NMPs is formed based on the distance from the main memory. In this paper, an evaluation tool is proposed to obtain the optimal design decision considering the power-time trade-off in the processor hierarchy. Every time the operating condition and constraints change, the decision of task-level offloading is dynamically made. For the realistic NMP-enabled system environment, the relationship among HBM, host, and NMP should be carefully considered. Hosts and NMPs are almost hidden from each other and the communications between them are extremely limited. In the simulation results, popular benchmarks and a machine learning application are used to demonstrate power-time trade-offs depending on applications and system conditions.

scholarly journals Reliability-aware Garbage Collection for Hybrid HBM-DRAM Memories

2021 ◽  
Vol 18 (1) ◽  
pp. 1-25
Author(s):  
Wenjie Liu ◽  
Shoaib Akram ◽  
Jennifer B. Sartor ◽  
Lieven Eeckhout
Keyword(s):  
Error Rate ◽  
Garbage Collection ◽  
Main Memory ◽  
Soft Error ◽  
Low Risk ◽  
Coarse Grained ◽  
Soft Error Rate ◽  
Page Migration ◽  
Trade Offs ◽  
High Bandwidth

Emerging workloads in cloud and data center infrastructures demand high main memory bandwidth and capacity. Unfortunately, DRAM alone is unable to satisfy contemporary main memory demands. High-bandwidth memory (HBM) uses 3D die-stacking to deliver 4–8× higher bandwidth. HBM has two drawbacks: (1) capacity is low, and (2) soft error rate is high. Hybrid memory combines DRAM and HBM to promise low fault rates, high bandwidth, and high capacity. Prior OS approaches manage HBM by mapping pages to HBM versus DRAM based on hotness (access frequency) and risk (susceptibility to soft errors). Unfortunately, these approaches operate at a coarse-grained page granularity, and frequent page migrations hurt performance. This article proposes a new class of reliability-aware garbage collectors for hybrid HBM-DRAM systems that place hot and low-risk objects in HBM and the rest in DRAM. Our analysis of nine real-world Java workloads shows that: (1) newly allocated objects in the nursery are frequently written, making them both hot and low-risk, (2) a small fraction of the mature objects are hot and low-risk, and (3) allocation site is a good predictor for hotness and risk. We propose RiskRelief, a novel reliability-aware garbage collector that uses allocation site prediction to place hot and low-risk objects in HBM. Allocation sites are profiled offline and RiskRelief uses heuristics to classify allocation sites as DRAM and HBM. The proposed heuristics expose Pareto-optimal trade-offs between soft error rate (SER) and execution time. RiskRelief improves SER by 9× compared to an HBM-Only system while at the same time improving performance by 29% compared to a DRAM-Only system. Compared to a state-of-the-art OS approach for reliability-aware data placement, RiskRelief eliminates all page migration overheads, which substantially improves performance while delivering similar SER. Reliability-aware garbage collection opens up a new opportunity to manage emerging HBM-DRAM memories at fine granularity while requiring no extra hardware support and leaving the programming model unchanged.

High-throughput Near-Memory Processing on CNNs with 3D HBM-like Memory

2021 ◽  
Vol 26 (6) ◽  
pp. 1-20
Author(s):  
Naebeom Park ◽  
Sungju Ryu ◽  
Jaeha Kung ◽  
Jae-Joon Kim
Keyword(s):  
Energy Consumption ◽  
High Performance ◽  
Three Dimensional ◽  
Cmos Technology ◽  
Memory Processing ◽  
Hybrid Memory ◽  
High Bandwidth ◽  
Dataflow Models ◽  
Accelerator Design ◽  
3D Memory

This article discusses the high-performance near-memory neural network (NN) accelerator architecture utilizing the logic die in three-dimensional (3D) High Bandwidth Memory– (HBM) like memory. As most of the previously reported 3D memory-based near-memory NN accelerator designs used the Hybrid Memory Cube (HMC) memory, we first focus on identifying the key differences between HBM and HMC in terms of near-memory NN accelerator design. One of the major differences between the two 3D memories is that HBM has the centralized through- silicon-via (TSV) channels while HMC has distributed TSV channels for separate vaults. Based on the observation, we introduce the Round-Robin Data Fetching and Groupwise Broadcast schemes to exploit the centralized TSV channels for improvement of the data feeding rate for the processing elements. Using synthesized designs in a 28-nm CMOS technology, performance and energy consumption of the proposed architectures with various dataflow models are evaluated. Experimental results show that the proposed schemes reduce the runtime by 16.4–39.3% on average and the energy consumption by 2.1–5.1% on average compared to conventional data fetching schemes.

An Energy-Efficient DRAM Cache Architecture for Mobile Platforms With PCM-Based Main Memory

2022 ◽  
Vol 21 (1) ◽  
pp. 1-22
Author(s):  
Dongsuk Shin ◽  
Hakbeom Jang ◽  
Kiseok Oh ◽  
Jae W. Lee
Keyword(s):  
Energy Consumption ◽  
Main Memory ◽  
Battery Life ◽  
Mobile Platforms ◽  
Total Energy Consumption ◽  
Efficient Manner ◽  
Hybrid Memory ◽  
Spatial Locality ◽  
Cache Architecture ◽  
Energy Delay Product

A long battery life is a first-class design objective for mobile devices, and main memory accounts for a major portion of total energy consumption. Moreover, the energy consumption from memory is expected to increase further with ever-growing demands for bandwidth and capacity. A hybrid memory system with both DRAM and PCM can be an attractive solution to provide additional capacity and reduce standby energy. Although providing much greater density than DRAM, PCM has longer access latency and limited write endurance to make it challenging to architect it for main memory. To address this challenge, this article introduces CAMP, a novel DRAM c ache a rchitecture for m obile platforms with P CM-based main memory. A DRAM cache in this environment is required to filter most of the writes to PCM to increase its lifetime, and deliver highest efficiency even for a relatively small-sized DRAM cache that mobile platforms can afford. To address this CAMP divides DRAM space into two regions: a page cache for exploiting spatial locality in a bandwidth-efficient manner and a dirty block buffer for maximally filtering writes. CAMP improves the performance and energy-delay-product by 29.2% and 45.2%, respectively, over the baseline PCM-oblivious DRAM cache, while increasing PCM lifetime by 2.7×. And CAMP also improves the performance and energy-delay-product by 29.3% and 41.5%, respectively, over the state-of-the-art design with dirty block buffer, while increasing PCM lifetime by 2.5×.

scholarly journals Sports training enhances visuo-spatial cognition regardless of open-closed typology

PeerJ ◽  
2017 ◽  
Vol 5 ◽  
pp. e3336 ◽  
Author(s):  
Ting-Yu Chueh ◽  
Chung-Ju Huang ◽  
Shu-Shih Hsieh ◽  
Kuan-Fu Chen ◽  
Yu-Kai Chang ◽  
...  
Keyword(s):  
Young Adults ◽  
Spatial Attention ◽  
Sport Participation ◽  
Memory Performance ◽  
Event Related Potential ◽  
Control Group ◽  
Attention And Memory ◽  
Sports Training ◽  
Memory Processing ◽  
Speed Accuracy

The aim of this study was to investigate the effects of open and closed sport participation on visuo-spatial attention and memory performance among young adults. Forty-eight young adults—16 open-skill athletes, 16 closed-skill athletes, and 16 non-athletes controls—were recruited for the study. Both behavioral performance and event-related potential (ERP) measurement were assessed when participants performed non-delayed and delayed match-to-sample task that tested visuo-spatial attention and memory processing. Results demonstrated that regardless of training typology, the athlete groups exhibited shorter reaction times in both the visuo-spatial attention and memory conditions than the control group with no existence of speed-accuracy trade-off. Similarly, a larger P3 amplitudes were observed in both athlete groups than in the control group for the visuo-spatial memory condition. These findings suggest that sports training, regardless of typology, are associated with superior visuo-spatial attention and memory performance, and more efficient neural resource allocation in memory processing.

APMigration: Improving Performance of Hybrid Memory Performance via An Adaptive Page Migration Method

2020 ◽  
Vol 31 (2) ◽  
pp. 266-278
Author(s):  
Yujuan Tan ◽  
Baiping Wang ◽  
Zhichao Yan ◽  
Witawas Srisa-an ◽  
Xianzhang Chen ◽  
...  
Keyword(s):  
Memory Performance ◽  
Hybrid Memory ◽  
Page Migration ◽  
Migration Method

scholarly journals The Application of Material Flow Cost Accounting in Waste Reduction

2019 ◽  
Vol 11 (5) ◽  
pp. 1270 ◽  
Author(s):  
Shaio Huang ◽  
An Chiu ◽  
Po Chao ◽  
Ni Wang
Keyword(s):  
Material Flow ◽  
Sustainable Management ◽  
Cost Accounting ◽  
Management Tool ◽  
Evaluation Tool ◽  
Use Of Resources ◽  
Primary Finding ◽  
Corporate Social ◽  
Effective Use ◽  
The Cost

Due to the rise in environmental awareness, corporate companies have shifted their focus from an obsession with short-term profits to contemplating long-term strategies to achieve sustainable management. Effective use of resources is the primary indicator of this achievement. Fulfillment of corporate social responsibility and thinking beyond the regulatory aspects of corporate sustainable management are goals that have continually attracted attention worldwide. Material flow cost accounting based on ISO 14051, which was announced by the International Organization for Standardization (ISO), is a tool that can be used to achieve a balance between the environment and economy. We focused on using ISO 14051-based material flow cost accounting as an analytical evaluation tool from the perspectives of finance and accounting personnel. We conducted a case study on a flat-panel parts supplier to determine whether the efficient use of recycled glass could reduce company costs. The primary finding is that the film layer on recycled washed glass tends to be stripped during the production process, causing increased reprocessing costs and thus rendering the cost of renewable cleaning higher than that of reworking. This study revealed that the ISO 14051-based material flow cost accounting analysis constitutes a valuable management tool, thereby facilitating the promotion of sustainable development.

Patient and clinician perspectives of an integrated electronic medication prescribing and dispensing system: A qualitative study at a multisite Australian hospital network

2017 ◽  
Vol 48 (1) ◽  
pp. 12-23 ◽  
Author(s):  
Grace Lau ◽  
Jayde Ho ◽  
Susan Lin ◽  
Karen Yeoh ◽  
Tiffany Wan ◽  
...  
Keyword(s):  
Positive Impact ◽  
Electronic Prescribing ◽  
Cross Sectional Survey ◽  
Cross Sectional ◽  
System A ◽  
Clinician Perspectives ◽  
Perceived Impact ◽  
Effective Use ◽  
Clinician Attitudes ◽  
Australian Hospital

Background: While clinician attitudes towards electronic prescribing (e-prescribing) systems have been widely studied, little is known about the perspectives of patients, despite being the primary beneficiaries of these systems. Objective: The objective of this study is to explore and compare patient and clinician attitudes towards an integrated e-prescribing and dispensing system, in order to guide improvements in system implementation, service delivery and enhancements to system functionality. Method: A cross-sectional survey was developed and administered to patients and multidisciplinary clinicians at a multisite Australian metropolitan teaching hospital network in all areas where e-prescribing was fully implemented. Participants’ views on perceived impact and valued features of the e-prescribing system were elucidated. Results: Overall, 783 participants (400 patients and 383 clinicians) completed the survey. Although 98% of clinicians were aware of the transition to e-prescriptions, only 36% of patients were aware prior to the study. Over 80% of patients and clinicians perceived improvements in prescribing and dispensing safety and clinician workflow; 90% of patients were comfortable with information privacy associated with e-prescriptions; and 86% of patients preferred e-prescriptions to handwritten prescriptions. Although over 80% of patients valued features that improved access to information and medication safety, clinicians were more discerning about valued system features. Conclusion: The majority of patients and clinicians reported a positive impact of e-prescribing on safety and efficiency. Both groups valued safe and effective use of medicines, although differences existed in the importance placed on key system features. A greater focus on patient engagement and communication is needed to optimise the delivery of patient-centred care.

scholarly journals Using Value Assessment to Drive PHM System Development in Early Design

2019 ◽  
Vol 11 (1) ◽  
Author(s):  
Daniel Hulse ◽  
Christopher Hoyle ◽  
Kai Goebel ◽  
Irem Tumer
Keyword(s):  
Health Management ◽  
Cost Model ◽  
Operational Performance ◽  
Design Decision ◽  
Value Assessment ◽  
Design Phase ◽  
Management Options ◽  
Early Design ◽  
Trade Offs ◽  
Early Design Phase

Prognostics and Health Management (PHM) systems have been shown to provide many benefits to the reliability, performance, and life of engineered systems. However, because of trade-offs between up-front design and implementation costs, operational performance, and reliability, it may not be obvious in the early design phase whether one PHM system will be more beneficial to another, or whether a PHM system will provide benefit compared to a traditional reliability approach. These trade-offs make the commitment required to pursue PHM features in the early design phase difficult to justify. In this paper, a cost model incorporating trade-offs among design cost, operational performance, and failure risk is used to provide a comprehensive value comparison of health management options to motivate design decision-making. This approach is then demonstrated in a simple case study comparing the use of a PHM system for condition-based maintenance or diagnostic-based recovery with implementing redundancy and increased inspection in the design. Then it is shown how different model inputs and assumptions result in a different system value (and different design choice from the process), illustrating the usefulness of cost modelling to capture design trade-offs. Using this approach, decisions about pursuing PHM can be made early, enabling the benefits to be fully leveraged in the design process to achieve increased operational resilience.

scholarly journals Medial prefrontal decoupling from the default mode network benefits memory

2019 ◽  
Author(s):  
N.C.J. Müller ◽  
M. Dresler ◽  
G. Janzen ◽  
C.F. Beckmann ◽  
G Fernández ◽  
...  
Keyword(s):  
Episodic Memory ◽  
Default Mode Network ◽  
Memory Performance ◽  
Task Demands ◽  
Parahippocampal Gyrus ◽  
Developmental Study ◽  
Cross Sectional ◽  
Memory Processing ◽  
Default Mode ◽  
New Material

AbstractIn the last few years the involvement of the medial prefrontal cortex (mPFC) in memory processing has received increased attention. It is centrally involved when we use prior knowledge (schemas) to improve learning of new material. With the mPFC also being one of the core hubs of the default mode network (DMN) and the DMN’s role in memory retrieval, we decided to investigate whether the mPFC in a schema paradigm acts independently of the DMN. We tested this with data from a cross-sectional developmental study. During retrieval of schema items, the mPFC decoupled from the DMN with the degree of decoupling predicting memory performance. This finding suggests that a demand specific reconfiguration of the DMN supports schema memory. Additionally, we found that in the control condition, which relied on episodic memory, activity in the parahippocampal gyrus was positively related to memory performance. We interpret these results as a demand specific network reconfiguration of the DMN: a decoupling of the mPFC to support schema memory and a decoupling of the parahippocampal gyrus facilitating episodic memory. This supports the notion of dynamic reconfiguration of brain networks in response to task demands in the sense of process specific alliances.

Sign in / Sign up

Export Citation Format

Share Document