arXiv subjects

Mingjun Zhang

Publications and source records attributed to Mingjun Zhang.

13 recordsLinked to original sources

arXiv2026

HetCCL: Enabling Collective Communication For Mixed-Vendor Heterogeneous Clusters

Training Large Language Models (LLMs) on heterogeneous clusters presents significant challenges for collective communication, as hardware from multiple vendors introduces diverse network and computational characteristics. Existing collective communication frameworks (e.g., NCCL, RCCL) designed for homogeneous environments fail to address mixed-hardware setups, while communication libraries with heterogeneous support (e.g., Gloo, OpenMPI) incur heavy overhead in the data path. This paper presents HetCCL, a framework that enables heterogeneous collective communication by efficient P2P transport across heterogeneous devices (e.g., GPUs), eliminating the host-device memory copy overhead while offloading the control to the CPUs. For combining collectives (e.g., AllReduce, ReduceScatter), HetCCL introduces a border-communicator mechanism that achieves vendor independence by using the intrinsic reduction in the combining collectives in vendor collective communication libraries. With efficient heterogeneous P2P transport and portable reduction mechanism, HetCCL proposes a hierarchical topology abstraction for heterogeneous clusters, dissecting collective communication into cluster-level primitives that guarantee optimal cross-cluster data transfer volume and optimal bandwidth utilization. We implement HetCCL with 4 different vendor support and evaluate it in 4 heterogeneous settings with benchmarks and end-to-end LLM tasks. Our evaluation shows that HetCCL achieves 17-19x higher bandwidth than Gloo in heterogeneous communications, and speeds up end-to-end training by up to 16.9% in the per-step-time.

cs.NI↗

arXiv2025

An Efficient, Reliable and Observable Collective Communication Library in Large-scale GPU Training Clusters

Large-scale LLM training requires collective communication libraries to exchange data among distributed GPUs. As a company dedicated to building and operating large-scale GPU training clusters, we encounter several practical limitations of NCCL in production, including 1) SM competition between computation and communication, 2) expensive restart costs under link failures, and 3) insufficient observability of transient collective communication anomalies. To address these challenges, we propose VCCL, an efficient, reliable, and observable collective communication library in large-scale GPU training clusters. VCCL removes SM-consuming P2P kernels by moving intra-node data movement and stream dependency enforcement to CPU threads and GPU copy engines. VCCL also introduces a primary-backup QP mechanism to tolerate frequent NIC port failures, and designs a window-based monitor to observe network anomalies at O({\mu}s) level. We opensource VCCL and deploy it in production training clusters for several months. Compared with NCCL, VCCL improves training throughput by up to 5.28% and reduces massive GPU resource wastage through runtime fault tolerance and finegrained monitor. We also share experience and lessons we learned during the deployment of VCCL in large-scale clusters.

cs.DC↗

arXiv2025

Expert-as-a-Service: Towards Efficient, Scalable, and Robust Large-scale MoE Serving

Mixture-of-Experts (MoE) models challenge serving infrastructures with dynamic, sparse expert utilization, causing instability on conventional systems designed for dense architectures. We propose EaaS, a novel serving system to enable efficient, scalable, and robust MoE deployment. Our system disaggregates MoE modules into independent, stateless services. This design enables fine-grained resource scaling and provides inherent fault tolerance by decoupling compute units. The architecture is powered by a high-performance, CPU-free peer-to-peer communication library that ensures minimal overhead and high throughput. Experiments confirm EaaS's scalability and efficiency, achieving performance comparable to monolithic systems while providing robust fault tolerance and strong scalability. EaaS incurs less than a 2% throughput reduction under simulated hardware failures that would otherwise halt monolithic architectures. It further saves up to 37.5% of computing resources through dynamic fine-grained adaptation to serving traffic, demonstrating strong resilience for large-scale MoE deployment in production.

cs.DC↗

arXiv2025

DistFlow: A Fully Distributed RL Framework for Scalable and Efficient LLM Post-Training

Effectively scaling Reinforcement Learning (RL) is crucial for enhancing the reasoning and alignment of Large Language Models. The massive data and complex execution flows inherent in these tasks require a distributed architecture capable of efficient scaling. However, to simplify programming and dependency management, mainstream frameworks often rely on a centralized architecture where a single node dispatches both control and data. This inherent coupling creates significant communication bottlenecks, severely limiting system scalability and efficiency. We present DISTFLOW, a novel, fully distributed RL framework that adopts a multi-controller paradigm. By decoupling data transmission from control dispatch, DISTFLOW establishes a parallelism-aware, decentralized Data Coordinator that leverages local caching, load balancing, and asynchronous double buffer to minimize communication overhead and mitigate straggler effects. For control logic, it introduces a task scheduler built upon Directed Acyclic Graph (DAG) that facilitates fine-grained, independent execution. Experimental results demonstrate that DISTFLOW achieves near-linear scalability up to 512 GPUs and delivers up to a 2.63x throughput improvement over state-of-the-art (SOTA) frameworks. The source code is available at: https://github.com/sii-research/siiRL.

cs.DC↗

arXiv2025

Achieving Stability and Optimality: Control Strategy for a Wind Turbine Supplying an Electrolyzer in the Islanded Storage-less Microgrid

Wind power generation supplying electrolyzers in islanded microgrids is an essential technical pathway for green hydrogen production, attracting growing attention in the transition towards net zero carbon emissions. Both academia and industry widely recognize that islanded AC microgrids normally rely on battery energy storage systems (BESSs) for grid-forming functions. However, the high cost of BESS significantly increases the levelized cost of hydrogen (LCOH), compromising economic feasibility. To address this challenge and reduce the LCOH, this paper focuses on a wind turbine (WT) supplying an electrolyzer in a storage-less microgrid and identifies a unique characteristic that challenges the conventional understanding of this microgrid: active power is coupled with microgrid voltage rather than frequency, the latter being entirely decoupled from active power balance. Based on this unique characteristic, this paper develops a new control strategy that maintains power balance, stabilizes the voltage and frequency, and maximizes hydrogen production. The effectiveness of the control strategy is validated through case studies conducted in Matlab/Simulink, especially its capability to maintain stability while maximizing hydrogen production under various conditions.

eess.SY↗

arXiv2024

Device Feature based on Graph Fourier Transformation with Logarithmic Processing For Detection of Replay Speech Attacks

The most common spoofing attacks on automatic speaker verification systems are replay speech attacks. Detection of replay speech heavily relies on replay configuration information. Previous studies have shown that graph Fourier transform-derived features can effectively detect replay speech but ignore device and environmental noise effects. In this work, we propose a new feature, the graph frequency device cepstral coefficient, derived from the graph frequency domain using a device-related linear transformation. We also introduce two novel representations: graph frequency logarithmic coefficient and graph frequency logarithmic device coefficient. We evaluate our methods using traditional Gaussian mixture model and light convolutional neural network systems as classifiers. On the ASVspoof 2017 V2, ASVspoof 2019 physical access, and ASVspoof 2021 physical access datasets, our proposed features outperform known front-ends, demonstrating their effectiveness for replay speech detection.

cs.SD↗

arXiv2023

Domain Similarity-Perceived Label Assignment for Domain Generalized Underwater Object Detection

The inherent characteristics and light fluctuations of water bodies give rise to the huge difference between different layers and regions in underwater environments. When the test set is collected in a different marine area from the training set, the issue of domain shift emerges, significantly compromising the model's ability to generalize. The Domain Adversarial Learning (DAL) training strategy has been previously utilized to tackle such challenges. However, DAL heavily depends on manually one-hot domain labels, which implies no difference among the samples in the same domain. Such an assumption results in the instability of DAL. This paper introduces the concept of Domain Similarity-Perceived Label Assignment (DSP). The domain label for each image is regarded as its similarity to the specified domains. Through domain-specific data augmentation techniques, we achieved state-of-the-art results on the underwater cross-domain object detection benchmark S-UODAC2020. Furthermore, we validated the effectiveness of our method in the Cityscapes dataset.

cs.CV↗

arXiv2022

New Techniques Based On Odd-Edge Total Colorings In Topological Cryptosystem

For building up twin-graphic lattices towards topological cryptograph, we define four kinds of new odd-magic-type colorings: odd-edge graceful-difference total coloring, odd-edge edge-difference total coloring, odd-edge edge-magic total coloring, and odd-edge felicitous-difference total coloring in this article. Our RANDOMLY-LEAF-ADDING algorithms are based on adding randomly leaves to graphs for producing continuously graphs admitting our new odd-magic-type colorings. We use complex graphs to make caterpillar-graphic lattices and complementary graphic lattices, such that each graph in these new graphic lattices admits a uniformly $W$-magic total coloring. On the other hands, finding some connections between graphic lattices and integer lattices is an interesting research, also, is important for application in the age of quantum computer. We set up twin-type $W$-magic graphic lattices (as public graphic lattices vs private graphic lattices) and $W$-magic graphic-lattice homomorphism for producing more complex topological number-based strings.

math.CO↗

arXiv2019

Topological Coding and Topological Matrices Toward Network Overall Security

A mathematical topology with matrix is a natural representation of a coding relational structure that is found in many fields of the world. Matrices are very important in computation of real applications, s ce matrices are easy saved in computer and run quickly, as well as matrices are convenient to deal with communities of current networks, such as Laplacian matrices, adjacent matrices in graph theory. Motivated from convenient, useful and powerful matrices used in computation and investigation of today's networks, we have introduced Topcode-matrices, which are matrices of order $3\times q$ and differ from popular matrices applied in linear algebra and computer science. Topcode-matrices can use numbers, letters, Chinese characters, sets, graphs, algebraic groups \emph{etc.} as their elements. One important thing is that Topcode-matrices of numbers can derive easily number strings, since number strings are text-based passwords used in information security. Topcode-matrices can be used to describe topological graphic passwords (Topsnut-gpws) used in information security and graph connected properties for solving some problems coming in the investigation of Graph Networks and Graph Neural Networks proposed by GoogleBrain and DeepMind. Our topics, in this article, are: Topsnut-matrices, Topcode-matrices, Hanzi-matrices, adjacency ve-value matrices and pan-Topcode-matrices, and some connections between these Topcode-matrices will be proven. We will discuss algebraic groups obtained from the above matrices, graph groups, graph networking groups and number string groups for encrypting different communities of dynamic networks. The operations and results on our matrices help us to set up our overall security mechanism to protect networks.

cs.IT↗

arXiv2019

Using Chinese Characters To Generate Text-Based Passwords For Information Security

Graphical passwords (GPWs) are in many areas of the current world. Topological graphic passwords (Topsnut-gpws) are a new type of cryptography, and they differ from the existing GPWs. A Topsnut-gpw consists of two parts: one is a topological structure (graph), and one is a set of discrete elements (a graph labelling, or coloring), the topological structure connects these discrete elements together to form an interesting "story". Our idea is to transform Chinese characters into computer and electronic equipments with touch screen by speaking, writing and keyboard for forming Hanzi-graphs and Hanzi-gpws. We will use Hanzigpws to produce text-based passwords (TB-paws). We will introduce flawed graph labellings on disconnected Hanzi-graphs.

cs.IT↗

arXiv2018

Text-based Passwords Generated From Topological Graphic Passwords

Topological graphic passwords (Topsnut-gpws) are one of graph-type passwords, but differ from the existing graphical passwords, since Topsnut-gpws are saved in computer by algebraic matrices. We focus on the transformation between text-based passwords (TB-paws) and Topsnut-gpws in this article. Several methods for generating TB-paws from Topsnut-gpws are introduced; these methods are based on topological structures and graph coloring/labellings, such that authentications must have two steps: one is topological structure authentication, and another is text-based authentication. Four basic topological structure authentications are introduced and many text-based authentications follow Topsnut-gpws. Our methods are based on algebraic, number theory and graph theory, many of them can be transformed into polynomial algorithms. A new type of matrices for describing Topsnut-gpws is created here, and such matrices can produce TB-paws in complex forms and longer bytes. Estimating the space of TB-paws made by Topsnut-gpws is very important for application. We propose to encrypt dynamic networks and try to face: (1) thousands of nodes and links of dynamic networks; (2) large numbers of Topsnut-gpws generated by machines rather than human's hands. As a try, we apply spanning trees of dynamic networks and graphic groups (Topsnut-groups) to approximate the solutions of these two problems. We present some unknown problems in the end of the article for further research.

cs.IT↗

arXiv2018

Topological Graphic Passwords And Their Matchings Towards Cryptography

Graphical passwords (GPWs) are convenient for mobile equipments with touch screen. Topological graphic passwords (Topsnut-gpws) can be saved in computer by classical matrices and run quickly than the existing GPWs. We research Topsnut-gpws by the matching of view, since they have many advantages. We discuss: configuration matching partition, coloring/labelling matching partition, set matching partition, matching chain, etc. And, we introduce new graph labellings for enriching Topsnut-matchings and show that these labellings can be realized for trees or spanning trees of networks. In theoretical works we explore Graph Labelling Analysis, and show that every graph admits our extremal labellings and set-type labellings in graph theory. Many of the graph labellings mentioned are related with problems of set matching partitions to number theory, and yield new objects and new problems to graph theory.

cs.CR↗

arXiv2010

Far-Field Plasmonic Resonance Enhanced Nano-Particle Image Velocimetry within a Micro Channel

In this paper, a novel far-field plasmonic resonance enhanced nanoparticle-seeded Particle Image Velocimetry (nPIV) has been demonstrated to measure the velocity profile in a micro channel. Chemically synthesized silver nanoparticles have been used to seed the flow in the micro channel. By using Discrete Dipole Approximation (DDA), plasmonic resonance enhanced light scattering has been calculated for spherical silver nanoparticles with diameters ranging from 15nm to 200nm. Optimum scattering wavelength is specified for the nanoparticles in two media: water and air. The diffraction-limited plasmonic resonance enhanced images of silver nanoparticles at different diameters have been recorded and analyzed. By using standard PIV techniques, the velocity profile within the micro channel has been determined from the images.

physics.flu-dyn↗