Xuan Zhang

Xuan Zhang

Auto-generated profile, based on publicly available information

Xuan Zhang

Northeastern University, Boston, MA, USA

Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 February 2020
Verified
Conference Paper
Closed access
UGUdit Gupta
Udit Gupta
CWCarole-Jean Wu
Carole-Jean Wu
XWXiaodong Wang
Xiaodong Wang
MNMaxim Naumov
Maxim Naumov
BRBrandon Reagen
Brandon Reagen
XZXuan Zhang
Xuan Zhang
The widespread application of deep learning has changed the landscape of computation in data centers. In particular, personalized recommendation for content ranking is now largely accomplished using deep neural networks. However, despite their importance and the amount of compute cycles they consume, relatively little research attention has been devoted to recommendation systems. To facilitate research and advance the understanding of these workloads, this paper presents a set of real-world, production-scale DNNs for personalized recommendation coupled with relevant performance metrics for evaluation. In addition to releasing a set of open-source workloads, we conduct in-depth analysis that underpins future system design and optimization for at-scale recommendation: Inference latency varies by 60% across three Intel server generations, batching and co-location of inference jobs can drastically improve latency-bounded throughput, and diversity across recommendation models leads to different optimization strategies.

16 July 2021 in IEEE Micro
Verified
Journal Article
Closed access
LKLiu Ke
Liu Ke
XZXuan Zhang
Xuan Zhang
JSJinin So
Jinin So
JLJong-Geon Lee
Jong-Geon Lee
SKShin-Haeng Kang
Shin-Haeng Kang
HLHsien-Hsin Sean Lee
Hsien-Hsin Sean Lee
Near-memory processing (NMP) is a prospective paradigm enabling memory-centric computing. By moving the compute capability next to the main memory (DRAM modules), it can fundamentally address the CPU-memory bandwidth bottleneck and thus effectively improve the performance of memory-constrained workloads. Using the personalized recommendation system as a driving example, we developed a scalable, practical DIMM-based NMP solution tailor-designed for accelerating the inference serving. Our solution is demonstrated on a versatile FPGA-enabled NMP platform called AxDIMM that allows rapid prototyping and evaluation of NMP's performance potential on real hardware under a realistic system setting using industry-representative recommendation framework. We experimentally validated the performance of a two-ranked AxDIMM prototype which achieves up to 1.89x speedup in latency and 31.6% memory energy saving for embedding operations. For end-to-end recommendation inference serving, AxDIMM improves the throughput up to 1.5x and latency-bounded throughput up to 1.77x, respectively.

1 November 2020 in Journal of Systems Architecture
Verified
Journal Article
Closed access
ABAdith Boloor
Adith Boloor
KGKarthik Garimella
Karthik Garimella
XHXin He
Xin He
CGChristopher Gill
Christopher Gill
YVYevgeniy Vorobeychik
Yevgeniy Vorobeychik
XZXuan Zhang
Xuan Zhang

17 March 2022 in Nature Nanotechnology
Verified
Journal Article
Open access
PDF file
WCWeidong Cao
Weidong Cao
CWChangqing Wang
Changqing Wang
WCWeijian Chen
Weijian Chen
SHSong Hu
Song Hu
HWHua Wang
Hua Wang
XZXuan Zhang
Xuan Zhang
Harnessing parity–time symmetry with balanced gain and loss profiles has created a variety of opportunities in electronics from wireless energy transfer to telemetry sensing and topological defect engineering. However, existing implementations often employ ad hoc approaches at low operating frequencies and are unable to accommodate large-scale integration. Here we report a fully integrated realization of parity–time symmetry in a standard complementary metal–oxide–semiconductor process technology. Our work demonstrates salient parity–time symmetry features such as phase transition as well as the ability to manipulate broadband microwave generation and propagation beyond the limitations encountered by existing schemes. The system shows 2.1 times the bandwidth and 30% noise reduction compared to conventional microwave generation in the oscillatory mode, and displays large non-reciprocal microwave transport from 2.75 to 3.10 GHz in the non-oscillatory mode due to enhanced nonlinearities. This approach could enrich integrated circuit design methodology beyond well-established performance limits and enable the use of scalable integrated circuit technology to study topological effects in high-dimensional non-Hermitian systems.

17 December 2010 in IEEE Transactions on Circuits and Systems I: Regular Papers
Verified
Journal Article
Closed access
XZXuan Zhang
Xuan Zhang
AAAlyssa B. Apsel
Alyssa B. Apsel
The design of a 1.8 GHz 3-stage current-starved ring oscillator with a process- and temperature- compensated current source is presented. Without post-fabrication calibration or off-chip components, the proposed low variation circuit is able to achieve a 65.1% reduction in the normalized standard deviation of its center frequency at room temperature and 85 ppm/ ° C temperature stability with no penalty in the oscillation frequency, the phase noise or the start-up time. Analysis on the impact of transistor scaling indicates that the same circuit topology can be applied to improve variability as feature size scales beyond the current deep submicron technology. Measurements taken on 167 test chips from two different lots fabricated in a standard 90 nm CMOS process show a 3x improvement in frequency variation compared to the baseline case of a conventional current-starved ring oscillator. The power and area for the proposed circuitry is 87 μW and 0.013 mm2 compared to 54 μ W and 0.01 mm 2 in the baseline case.

12 September 2019 in IEEE Transactions on Circuits and Systems I: Regular Papers
Verified
Journal Article
Closed access
ZCZhe Chen
Zhe Chen
HZHuifeng Zhu
Huifeng Zhu
ERErxiang Ren
Erxiang Ren
ZLZheyu Liu
Zheyu Liu
KJKaige Jia
Kaige Jia
HYHuazhong Yang
Huazhong Yang
In the era of Artificial Intelligence (AI), bio-inspired perceptual computing system design brings favorable opportunities, while still facing considerable challenges in the meantime. Especially for tasks of image recognition in power-limited vision-based Internet of Things (IoT) devices, energy constraints due to the end of Dennard scaling limit the performance of Neural Network (NN) algorithms on popular digital platforms, which would not reach the energy efficiency requirement for embedded AI applications. In this paper, a processing near sensor architecture in mixed-signal domain with CMOS Image Sensor (CIS) of convolutional-kernel-readout method is proposed. Visual data is collected from a smart CIS, which can realize maximum $5 x5$ kernel-readout with minimum one slide step for convolutional operations. The outputs of CIS are directly processed by analog processing units locating near CIS without the constraint of digital clock and bottleneck of Analog-to-Digital Converter (ADC). By analyzing the effects of analog noise on classification accuracy, we further evaluate the fault-tolerance of the system to circuit noise and the device imperfection, such as mismatch and process variation. A mixed-signal visual perception chip is fabricated with a 32 x 32 image sensor and a Binarized Neural Network (BNN) processing array integrated with SMIC 180nm standard CMOS mixed-signal process. Measurement results show up to 545.4 GOPS/W energy efficiency with 1.8mW power consumption taking the advantages of ADC-free processing architecture. This work provides a promising alternative for low-power vision-based IoT intelligent applications.