Jie Li

Mrs. Jie Li


Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
10 May 2023
Verified
Book Chapter
Open access
PDF file
JLJie Li
Jie Li
GMGeorge Michelogiannakis
George Michelogiannakis
BCBrandon Cook
Brandon Cook
DCDulanya Cooray
Dulanya Cooray
YCYong Chen
Yong Chen
Resource demands of HPC applications vary significantly. However, it is common for HPC systems to primarily assign resources on a per-node basis to prevent interference from co-located workloads. This gap between the coarse-grained resource allocation and the varying resource demands can lead to HPC resources being not fully utilized. In this study, we analyze the resource usage and application behavior of NERSC’s Perlmutter, a state-of-the-art open-science HPC system with both CPU-only and GPU-accelerated nodes. Our one-month usage analysis reveals that CPUs are commonly not fully utilized, especially for GPU-enabled jobs. Also, around 64% of both CPU and GPU-enabled jobs used 50% or less of the available host memory capacity. Additionally, about 50% of GPU-enabled jobs used up to 25% of the GPU memory, and the memory capacity was not fully utilized in some ways for all jobs. While our study comes early in Perlmutter’s lifetime thus policies and application workload may change, it provides valuable insights on performance characterization, application behavior, and motivates systems with more fine-grain resource allocation.

18 November 2022 in Applied Mathematical Modelling
Verified
Journal Article
Closed access
SDShuangli Du
Shuangli Du
MZMinghua Zhao
Minghua Zhao
YLYiguang Liu
Yiguang Liu
ZYZhenzhen You
Zhenzhen You
Zhenghao ShiZhenghao Shi
Zhenghao Shi
ZXZhenyu Xu
Zhenyu Xu

18 March 2021 in IEEE Transactions on Computers
Verified
Journal Article
Closed access
XWXi Wang
Xi Wang
ATAntonino Tumeo
Antonino Tumeo
JLJie Li
Jie Li
JLJie Li
Jie Li
YCYong Chen
Yong Chen
Emerging High-Performance Computing (HPC) workloads, such as graph analytics, machine learning, and big data science, are data-intensive. Data-intensive workloads usually present fine-grained memory accesses with limited or no data locality, and thus incur frequent cache misses and low utilization of memory bandwidth. 3D-stacked memory devices such as Hybrid Memory Cube (HMC) and High Bandwidth Memory (HBM) can provide significantly higher bandwidth than conventional memory modules. However, the traditional interfaces and optimization methods for JEDEC DDR devices do not allow to fully exploit the potential performance of 3D-stacked memory with the massive amount of irregular memory accesses of data-intensive applications. In this article, we propose a novel Hotspot-Aware Manager (HAM) infrastructure for 3D-stacked memory devices capable of optimizing memory access streams via request aggregation, hotspot detection, and in-memory prefetching. We present the HAM design and implementation, and simulate it on a system using RISC-V embedded cores with attached HMC devices. We extensively evaluate HAM with over 12 benchmarks and applications representing diverse irregular memory access patterns. The results show that, on average, HAM reduces redundant requests by 37.51 percent and increases the prefetch buffer hit rate by 4.2 times, compared to a baseline streaming prefetcher. On the selected benchmark set, HAM provides performance gains of 21.81 percent in average (up to 34.28 percent), and power savings of 35.07 percent over a standard 3D-stacked memory.

26 July 2020
Verified
Conference Paper
Open access
NNNgan Nguyen
Ngan Nguyen
JHJon Hass
Jon Hass
YCYong Chen
Yong Chen
JLJie Li
Jie Li
ASAlan Sill
Alan Sill
TDTommy Dang
Tommy Dang
This showcase presents a visual approach based on clustering and superimposing to construct a high-level overview of sequential event data while balancing the amount of information and the cardinality in it. We also implement an interactive prototype, called RadarViewer , that allows domain analysts to simultaneously analyze sequence clustering, extract useful distribution patterns, drill multiple levels-of-detail to accelerate the analysis. The RadarViewer is demonstrated through case studies with real-world temporal datasets of different sizes.

4 August 2022
Verified
Conference Paper
Closed access
PDF file
JLJie Li
Jie Li
XZXiaosong Zhang
Xiaosong Zhang
JYJunkai Yi
Junkai Yi
RCRuiRong Chen
RuiRong Chen
In the current big data environment, aiming at the problems that traditional machine learning needs manual intervention and time-consuming to detect DDoS attacks, a DDoS attack detection method based on a double-stacked long short-term memory network is presented. The preprocessed data stream is sorted by recursive feature elimination algorithm, and the features with the most DDoS attack characteristics are selected as high-quality features, forming a double-stacked long short-term memory network data input format. The Center Loss is introduced into the Softmax Loss to reduce the intraclass distance, further improve the classification accuracy. Finally, the information containing DDoS attack characteristics can be quickly extracted from the complex characteristics of traffic. The CIC-IDS2017 dataset is used to train the model. Experimental research shows that the proposed model has an accuracy rate of 99.48% compared with other neural network models, and the detection effect is better than the compared algorithms, which can effectively achieve DDoS attack detection.

2 December 2025 in arXiv
Verified
Preprint
Open access
PDF file
CNChenxu Niu
Chenxu Niu
WZWei Zhang
Wei Zhang
JLJie Li
Jie Li
YZYongjian Zhao
Yongjian Zhao
TWTongyang Wang
Tongyang Wang
YCYong Chen
Yong Chen
Large language model (LLM) services now answer billions of queries per day, and industry reports show that inference, not training, accounts for more than 90% of total power consumption. However, existing benchmarks focus on either training/fine-tuning or performance of inference and provide little support for power consumption measurement and analysis of inference. We introduce TokenPowerBench, the first lightweight and extensible benchmark designed for LLM-inference power consumption studies. The benchmark combines (i) a declarative configuration interface covering model choice, prompt set, and inference engine, (ii) a measurement layer that captures GPU-, node-, and system-level power without specialized power meters, and (iii) a phase-aligned metrics pipeline that attributes energy to the prefill and decode stages of every request. These elements make it straight-forward to explore the power consumed by an LLM inference run; furthermore, by varying batch size, context length, parallelism strategy and quantization, users can quickly assess how each setting affects joules per token and other energy-efficiency metrics. We evaluate TokenPowerBench on four of the most widely used model series (Llama, Falcon, Qwen, and Mistral). Our experiments cover from 1 billion parameters up to the frontier-scale Llama3-405B model. Furthermore, we release TokenPowerBench as open source to help users to measure power consumption, forecast operating expenses, and meet sustainability targets when deploying LLM services.