the proceedings contain 53 papers. the topics discussed include: Nomad: non-exclusive memory tiering via transactional page migration;managing memory tiers with CXL in virtualized environments;harvesting memory-bound ...
ISBN:
(纸本)9781939133403
the proceedings contain 53 papers. the topics discussed include: Nomad: non-exclusive memory tiering via transactional page migration;managing memory tiers with CXL in virtualized environments;harvesting memory-bound CPU stall cycles in software with MSH;a tale of two paths: toward a hybrid data plane for efficient far-memory applications;DRust: language-guided distributed shared memory with fine granularity, full transparency, and ultra efficiency;taming throughput-latency tradeoff in LLM inference with Sarathi-Serve;ServerlessLLM: low-latency serverless inference for large language models;InfiniGen: efficient generative inference of large language models with dynamic KV cache management;Llumnix: dynamic scheduling for large language model serving;DistServe: disaggregating prefill and decoding for goodput-optimized large language model serving;and beaver: practical partial snapshots for distributed cloud services.
the proceedings contain 59 papers. the topics discussed include: accessing cloud with disaggregated software-defined router;CodedBulk: inter-datacenter bulk transfers using network coding;twenty years after: hierarchi...
ISBN:
(纸本)9781939133212
the proceedings contain 59 papers. the topics discussed include: accessing cloud with disaggregated software-defined router;CodedBulk: inter-datacenter bulk transfers using network coding;twenty years after: hierarchical core-stateless fair queuing;breaking the transience-equilibrium nexus: a new approach to datacenter packet transport;Orion: Google’s software-defined networking control plane;finding invariants of distributed systems: it’s a small (enough) world after all;Avenir: managing data plane diversity with control plane synthesis;contracting wide-area network topologies to solve flow problems quickly;cost-effective cloud edge traffic engineering with CASCARA;and a social network under social distancing: risk-driven backbone management during COVID-19 and beyond.
the study explores implementation of occupant behavior(OB) models during Conceptual design stage of a development project in the building industry. Despite extensive research in this field, there has been a notable ga...
详细信息
With smart grids replacing traditional energy grids in recent years, energy companies are facing a new challenge to efficiently manage and analyze the collected data. the use of sensors in smart grids has led to a sur...
详细信息
ISBN:
(纸本)9798350358810;9798350358803
With smart grids replacing traditional energy grids in recent years, energy companies are facing a new challenge to efficiently manage and analyze the collected data. the use of sensors in smart grids has led to a surge in the data being collected. Due to its design, the classic relational Online Transaction Processing (OLTP) relational database management system (RDBMS) starts to become inefficient for queries against the RDBMS beyond a terabyte size. For data extraction efficiency on large volumes of data, OLAP along with data warehousing (DW) has become a popular solution. A Business Intelligence (BI) tools are often used on the top of DWs [1]. the adoption of big-data-driven technologies is still lagging among energy companies. In this paper, based on our previous research projects and research results [2]-[5], we show that, through an optimization process, a gain in performance that is up to 2800 times faster compared withthat of the original OLTP DBMS using the same hardware and the same operating system is possible. the applied research project was funded by an NSERC grant and was conducted in 2020-2023 at Okanagan and Langara Colleges.
In this paper, we present a system-on-chip (SoC) design for incorrect face mask-wearing detection on a low-cost FPGA device. the SoC architecture realizes a VGG9 model to classify the incorrect face mask-wearing perso...
详细信息
the proceedings contain 100 papers. the topics discussed include: hybrid energy storage system to increase energy efficiency during regenerative braking of urban electrical transport;bidirectional flying inductor conv...
ISBN:
(纸本)9798350311273
the proceedings contain 100 papers. the topics discussed include: hybrid energy storage system to increase energy efficiency during regenerative braking of urban electrical transport;bidirectional flying inductor converter based on Karschny topology;feasibility study of a hybrid excited claw-pole alternator;impact of decarbonization in the energy sector on low-voltage distribution networks;operational torque characteristics of innovative coaxial magnetic gear's constructions;inverter model optimization in Simulink software environment;classification of factors affecting employee satisfaction in modern business organizations;implementation of plc-based cooling system control of an industrial hall where a large amount of heat is emitted;design for sustainable behavior: a focus on sustainable energy;and analytical and simulation study of the power losses of a switched reluctance motor.
this research investigates the development of a rectangular microstrip antenna array, employing an aperture coupled patch feeding technique, specifically designed to facilitate communication between ground stations an...
详细信息
design and implementation of real care processes is a complex task, where the information model is central and serves as a basis for further implementation of the overall IT system. In this contribution, we describe t...
详细信息
this paper presents the design and implementation of a 5-bit 18.5 GS/s flash ADC. the focus of the work has been on the single-channel implementation of the ADC at the target speed to avoid extra complexity and effort...
详细信息
ISBN:
(纸本)9782874870736
this paper presents the design and implementation of a 5-bit 18.5 GS/s flash ADC. the focus of the work has been on the single-channel implementation of the ADC at the target speed to avoid extra complexity and effort introduced by time-interleaved converters. this flash ADC according to our best knowledge is the fastest single-core 5-bit flash ever realized in CMOS technology. Utilization of a two-stage latch design along with forward-body-biasing (FBB) are the key points to achieve the required comparator speed. Furthermore, employing a novel body-biased preamplifier to generate the flash references, the power-hungry low-ohmic R-ladder and the classical differential difference amplifier (DDA) are eliminated saving huge power as well as reducing the kickback noise. the SF-based T&H core operates from a 1.2V supply consuming 67mW. the rest of the core blocks including clock-trees use a 0.9V supply and consume 200mW. Characterization results reveal a Nyquist SFDR and SNDR of 33.6 dBc and 25 dB (ENOB=3.9 bits) respectively using a full-scale single-tone sine wave input (f(in) = 9.1GHz). the LF SFDR and SNDR are 35.2 dBc and 28 dB (ENOB=4.3 bits) respectively (f(in) = 4.7MHz). the core ADC with pads occupies an area of 1.01mm(2), while the active area is 0.06mm(2).
this paper presents a novel capacitive sensor designed for real-time monitoring of the infusion drip rate during intravenous (IV) therapy administration. the proposed sensor, operating on the principle of proximity ca...
详细信息
暂无评论