arXiv subjects

Xiangke Liao

Publications and source records attributed to Xiangke Liao.

18 recordsLinked to original sources

arXiv2026

Physically Partitioned KVCache Format for CPU--GPU Load Balancing in MoE Inference

Single-GPU long-context inference with Mixture-of-Experts (MoE) models requires spilling the key-value cache (KVCache) to CPU memory. The spilled KV serves two complementary purposes---transferring to the GPU for attention computation, or computing in-place on the CPU---which demand opposing physical states. The optimal split between them varies with workload, yet existing KVCache abstractions offer only storage semantics over a monolithic object of a single physical state, and cannot express dynamic load balancing. We propose InplaceKVCache, the first KVCache abstraction whose format fixes each byte's physical residency at write time, so that the CPU--GPU load balance can be adjusted without moving data after placement. It realizes this as a four-region layout along two dimensions---device affinity and access pattern---turning load balancing into pure scheduling. Built on this abstraction, WriteScope splits CPU--GPU shares along the sequence dimension, and a portable roofline performance model determines the optimal CPU share as sequence length evolves, with online feedback tracking CPU cost drift. On three MoE models (DeepSeek-V2-Lite, Qwen3-30B-A3B, Mixtral-8$\times$7B) with a 32~GB VRAM budget, WriteScope supports end-to-end inference at the 1M-token aggregate scale. In the long-context regime ($\ge$8K), it achieves geometric-mean speedups of $1.5\times$--$2.5\times$ on A100 and $1.4\times$--$1.7\times$ on V100 over four reproduced baselines, while vLLM, SGLang, and KTransformers fail even with a doubled KV budget. A DeepSeek-V4-Flash case study validates composition with native sparse attention.

cs.DC↗

arXiv2026

LayerScope: Predictive Cross-Layer Scheduling for Efficient Multi-Batch MoE Inference on Legacy Servers

Mixture-of-Experts (MoE) models face memory and PCIe latency bottlenecks when deployed on commodity hardware. Offloading expert weights to CPU memory results in PCIe transfer latency that exceeds GPU computation by several folds. We present PreScope, a prediction-driven expert scheduling system that addresses three key challenges: inaccurate activation prediction, PCIe bandwidth competition, and cross-device scheduling complexity. Our solution includes: 1) Learnable Layer-Aware Predictor (LLaPor) that captures layer-specific expert activation patterns; 2) Prefetch-Aware Cross-Layer Scheduling (PreSched) that generates globally optimal plans balancing prefetching costs and loading overhead; 3) Asynchronous I/O Optimizer (AsyncIO) that decouples I/O from computation, eliminating waiting bubbles. PreScope achieves 141% higher throughput and 74.6% lower latency than state-of-the-art solutions.

cs.LG↗

arXiv2026

Nezha: Breaking Multi-Rail Network Barriers for Distributed DNN Training

In distributed deep learning, communication remains a critical bottleneck. While modern hardware advances rapidly, over 60 percent of production HPC systems still rely on legacy infrastructure (V100 GPUs, multi-plane Ethernet/InfiniBand), necessitating communication optimization without hardware upgrades. Existing approaches face three key limitations: (1) static single-rail binding underutilizes multi-rail bandwidth, (2) protocol heterogeneity (TCP-RDMA coexistence) causes synchronization delays, and (3) mainstream libraries (NCCL/MPI) lack cross-protocol coordination. We present Nezha, the first protocol-agnostic system for multi-rail networks. Our contributions include: (1) Hardware-agnostic cross-protocol coordination: A unified abstraction enabling seamless collaboration between in-network computing (SHARP), adaptive RDMA (GLEX), and TCP, achieving 1.7 to 4.3 times lower latency than Gloo. (2) Protocol-aware dynamic load balancing: A hybrid scheduling strategy with cold/hot start state machine for heterogeneous protocols, reducing startup latency for small payloads while enhancing throughput for large transfers. (3) Fault-tolerant multi-rail collaboration: A self-recovery mechanism that reroutes data flows within 200 milliseconds upon single-rail failures, ensuring uninterrupted training. Experiments on 8-node clusters demonstrate Nezha achieves 74 percent and 80 percent higher throughput than MPTCP in homogeneous (TCP-TCP) and heterogeneous (TCP-SHARP) networks, respectively. On 128-node supercomputers, Nezha delivers 2.36 times higher training efficiency than Gloo. By bridging modern DNN communication demands with legacy infrastructure, Nezha proves that systematic multi-rail optimization can unlock the potential of aging clusters.

cs.DC↗

arXiv2025

Unseen Horizons: Unveiling the Real Capability of LLM Code Generation Beyond the Familiar

Recently, large language models (LLMs) have shown strong potential in code generation tasks. However, there are still gaps before they can be fully applied in actual software development processes. Accurately assessing the code generation capabilities of large language models has become an important basis for evaluating and improving the models. Some existing works have constructed datasets to evaluate the capabilities of these models. However, the current evaluation process may encounter the illusion of "Specialist in Familiarity", primarily due to three gaps: the exposure of target code, case timeliness, and dependency availability. The fundamental reason for these gaps is that the code in current datasets may have been extensively exposed and exercised during the training phase, and due to the continuous training and development of LLM, their timeliness has been severely compromised. The key to solve the problem is to, as much as possible, evaluate the LLMs using code that they have not encountered before. Thus, the fundamental idea in this paper is to draw on the concept of code obfuscation, changing code at different levels while ensuring the functionality and output. To this end, we build a code-obfuscation based benchmark OBFUSEVAL. We first collect 1,354 raw cases from five real-world projects, including function description and code. Then we use three-level strategy (symbol, structure and semantic) to obfuscate descriptions, code and context dependencies. We evaluate four LLMs on OBFU- SEVAL and compared the effectiveness of different obfuscation strategy. We use official test suites of these projects to evaluate the generated code. The results show that after obfuscation, the average decrease ratio of test pass rate can up to 62.5%.

cs.SE↗

arXiv2023

Bridging Code Semantic and LLMs: Semantic Chain-of-Thought Prompting for Code Generation

Large language models (LLMs) have showcased remarkable prowess in code generation. However, automated code generation is still challenging since it requires a high-level semantic mapping between natural language requirements and codes. Most existing LLMs-based approaches for code generation rely on decoder-only causal language models often treate codes merely as plain text tokens, i.e., feeding the requirements as a prompt input, and outputing code as flat sequence of tokens, potentially missing the rich semantic features inherent in source code. To bridge this gap, this paper proposes the "Semantic Chain-of-Thought" approach to intruduce semantic information of code, named SeCoT. Our motivation is that the semantic information of the source code (\eg data flow and control flow) describes more precise program execution behavior, intention and function. By guiding LLM consider and integrate semantic information, we can achieve a more granular understanding and representation of code, enhancing code generation accuracy. Meanwhile, while traditional techniques leveraging such semantic information require complex static or dynamic code analysis to obtain features such as data flow and control flow, SeCoT demonstrates that this process can be fully automated via the intrinsic capabilities of LLMs (i.e., in-context learning), while being generalizable and applicable to challenging domains. While SeCoT can be applied with different LLMs, this paper focuses on the powerful GPT-style models: ChatGPT(close-source model) and WizardCoder(open-source model). The experimental study on three popular DL benchmarks (i.e., HumanEval, HumanEval-ET and MBPP) shows that SeCoT can achieves state-of-the-art performance, greatly improving the potential for large models and code generation.

cs.CL↗

arXiv2023

Learning to Branch in Combinatorial Optimization with Graph Pointer Networks

Branch-and-bound is a typical way to solve combinatorial optimization problems. This paper proposes a graph pointer network model for learning the variable selection policy in the branch-and-bound. We extract the graph features, global features and historical features to represent the solver state. The proposed model, which combines the graph neural network and the pointer mechanism, can effectively map from the solver state to the branching variable decisions. The model is trained to imitate the classic strong branching expert rule by a designed top-k Kullback-Leibler divergence loss function. Experiments on a series of benchmark problems demonstrate that the proposed approach significantly outperforms the widely used expert-designed branching rules. Our approach also outperforms the state-of-the-art machine-learning-based branch-and-bound methods in terms of solving speed and search tree size on all the test instances. In addition, the model can generalize to unseen instances and scale to larger instances.

cs.LG↗

arXiv2023

Large Language Models are Few-Shot Summarizers: Multi-Intent Comment Generation via In-Context Learning

Code comment generation aims at generating natural language descriptions for a code snippet to facilitate developers' program comprehension activities. Despite being studied for a long time, a bottleneck for existing approaches is that given a code snippet, they can only generate one comment while developers usually need to know information from diverse perspectives such as what is the functionality of this code snippet and how to use it. To tackle this limitation, this study empirically investigates the feasibility of utilizing large language models (LLMs) to generate comments that can fulfill developers' diverse intents. Our intuition is based on the facts that (1) the code and its pairwise comment are used during the pre-training process of LLMs to build the semantic connection between the natural language and programming language, and (2) comments in the real-world projects, which are collected for the pre-training, usually contain different developers' intents. We thus postulate that the LLMs can already understand the code from different perspectives after the pre-training. Indeed, experiments on two large-scale datasets demonstrate the rationale of our insights: by adopting the in-context learning paradigm and giving adequate prompts to the LLM (e.g., providing it with ten or more examples), the LLM can significantly outperform a state-of-the-art supervised learning approach on generating comments with multiple intents. Results also show that customized strategies for constructing the prompts and post-processing strategies for reranking the results can both boost the LLM's performances, which shed light on future research directions for using LLMs to achieve comment generation.

cs.SE↗

arXiv2023

One Adapter for All Programming Languages? Adapter Tuning for Code Search and Summarization

As pre-trained models automate many code intelligence tasks, a widely used paradigm is to fine-tune a model on the task dataset for each programming language. A recent study reported that multilingual fine-tuning benefits a range of tasks and models. However, we find that multilingual fine-tuning leads to performance degradation on recent models UniXcoder and CodeT5. To alleviate the potentially catastrophic forgetting issue in multilingual models, we fix all pre-trained model parameters, insert the parameter-efficient structure adapter, and fine-tune it. Updating only 0.6\% of the overall parameters compared to full-model fine-tuning for each programming language, adapter tuning yields consistent improvements on code search and summarization tasks, achieving state-of-the-art results. In addition, we experimentally show its effectiveness in cross-lingual and low-resource scenarios. Multilingual fine-tuning with 200 samples per programming language approaches the results fine-tuned with the entire dataset on code summarization. Our experiments on three probing tasks show that adapter tuning significantly outperforms full-model fine-tuning and effectively overcomes catastrophic forgetting.

cs.SE↗

arXiv2022

SAIH: A Scalable Evaluation Methodology for Understanding AI Performance Trend on HPC Systems

Novel artificial intelligence (AI) technology has expedited various scientific research, e.g., cosmology, physics and bioinformatics, inevitably becoming a significant category of workload on high performance computing (HPC) systems. Existing AI benchmarks tend to customize well-recognized AI applications, so as to evaluate the AI performance of HPC systems under predefined problem size, in terms of datasets and AI models. Due to lack of scalability on the problem size, static AI benchmarks might be under competent to help understand the performance trend of evolving AI applications on HPC systems, in particular, the scientific AI applications on large-scale systems. In this paper, we propose a scalable evaluation methodology (SAIH) for analyzing the AI performance trend of HPC systems with scaling the problem sizes of customized AI applications. To enable scalability, SAIH builds a set of novel mechanisms for augmenting problem sizes. As the data and model constantly scale, we can investigate the trend and range of AI performance on HPC systems, and further diagnose system bottlenecks. To verify our methodology, we augment a cosmological AI application to evaluate a real HPC system equipped with GPUs as a case study of SAIH.

cs.DC↗

arXiv2022

Bridging Pre-trained Models and Downstream Tasks for Source Code Understanding

With the great success of pre-trained models, the pretrain-then-finetune paradigm has been widely adopted on downstream tasks for source code understanding. However, compared to costly training a large-scale model from scratch, how to effectively adapt pre-trained models to a new task has not been fully explored. In this paper, we propose an approach to bridge pre-trained models and code-related tasks. We exploit semantic-preserving transformation to enrich downstream data diversity, and help pre-trained models learn semantic features invariant to these semantically equivalent transformations. Further, we introduce curriculum learning to organize the transformed data in an easy-to-hard manner to fine-tune existing pre-trained models. We apply our approach to a range of pre-trained models, and they significantly outperform the state-of-the-art models on tasks for source code understanding, such as algorithm classification, code clone detection, and code search. Our experiments even show that without heavy pre-training on code data, natural language pre-trained model RoBERTa fine-tuned with our lightweight approach could outperform or rival existing code pre-trained models fine-tuned on the above tasks, such as CodeBERT and GraphCodeBERT. This finding suggests that there is still much room for improvement in code pre-trained models.

cs.SE↗

arXiv2021

deGraphCS: Embedding Variable-based Flow Graph for Neural Code Search

With the rapid increase in the amount of public code repositories, developers maintain a great desire to retrieve precise code snippets by using natural language. Despite existing deep learning based approaches(e.g., DeepCS and MMAN) have provided the end-to-end solutions (i.e., accepts natural language as queries and shows related code fragments retrieved directly from code corpus), the accuracy of code search in the large-scale repositories is still limited by the code representation (e.g., AST) and modeling (e.g., directly fusing the features in the attention stage). In this paper, we propose a novel learnable deep Graph for Code Search (calleddeGraphCS), to transfer source code into variable-based flow graphs based on the intermediate representation technique, which can model code semantics more precisely compared to process the code as text directly or use the syntactic tree representation. Furthermore, we propose a well-designed graph optimization mechanism to refine the code representation, and apply an improved gated graph neural network to model variable-based flow graphs. To evaluate the effectiveness of deGraphCS, we collect a large-scale dataset from GitHub containing 41,152 code snippets written in C language, and reproduce several typical deep code search methods for comparison. Besides, we design a qualitative user study to verify the practical value of our approach. The experimental results have shown that deGraphCS can achieve state-of-the-art performances, and accurately retrieve code snippets satisfying the needs of the users.

cs.SE↗

arXiv2021

CD-SGD: Distributed Stochastic Gradient Descent with Compression and Delay Compensation

Communication overhead is the key challenge for distributed training. Gradient compression is a widely used approach to reduce communication traffic. When combining with parallel communication mechanism method like pipeline, gradient compression technique can greatly alleviate the impact of communication overhead. However, there exists two problems of gradient compression technique to be solved. Firstly, gradient compression brings in extra computation cost, which will delay the next training iteration. Secondly, gradient compression usually leads to the decrease of convergence accuracy.

cs.LG↗

arXiv2021

SSD-SSD: Communication sparsification for distributed deep learning training

Intensive communication and synchronization cost for gradients and parameters is the well-known bottleneck of distributed deep learning training. Based on the observations that Synchronous SGD (SSGD) obtains good convergence accuracy while asynchronous SGD (ASGD) delivers a faster raw training speed, we propose Several Steps Delay SGD (SSD-SGD) to combine their merits, aiming at tackling the communication bottleneck via communication sparsification. SSD-SGD explores both global synchronous updates in the parameter servers and asynchronous local updates in the workers in each periodic iteration. The periodic and flexible synchronization makes SSD-SGD achieve good convergence accuracy and fast training speed. To the best of our knowledge, we strike the new balance between synchronization quality and communication sparsification, and improve the trade-off between accuracy and training speed. Specifically, the core components of SSD-SGD include proper warm-up stage, steps delay stage, and our novel algorithm of global gradient for local update (GLU). GLU is critical for local update operations to effectively compensate the delayed local weights. Furthermore, we implement SSD-SGD on MXNet framework and comprehensively evaluate its performance with CIFAR-10 and ImageNet datasets. Experimental results show that SSD-SGD can accelerate distributed training speed under different experimental configurations, by up to 110%, while achieving good convergence accuracy.

cs.DC↗

arXiv2021

ConfInLog: Leveraging Software Logs to Infer Configuration Constraints

Misconfigurations have become the dominant causes of software failures in recent years, drawing tremendous attention for their increasing prevalence and severity. Configuration constraints can preemptively avoid misconfiguration by defining the conditions that configuration options should satisfy. Documentation is the main source of configuration constraints, but it might be incomplete or inconsistent with the source code. In this regard, prior researches have focused on obtaining configuration constraints from software source code through static analysis. However, the difficulty in pointer analysis and context comprehension prevents them from collecting accurate and comprehensive constraints. In this paper, we observed that software logs often contain configuration constraints. We conducted an empirical study and summarized patterns of configuration-related log messages. Guided by the study, we designed and implemented ConfInLog, a static tool to infer configuration constraints from log messages. ConfInLog first selects configuration-related log messages from source code by using the summarized patterns, then infers constraints from log messages based on the summarized natural language patterns. To evaluate the effectiveness of ConfInLog, we applied our tool on seven popular open-source software systems. ConfInLog successfully inferred 22 to 163 constraints, in which 59.5% to 61.6% could not be inferred by the state-of-the-art work. Finally, we submitted 67 documentation patches regarding the constraints inferred by ConfInLog. The constraints in 29 patches have been confirmed by the developers, among which 10 patches have been accepted.

cs.SE↗

arXiv2021

DepOwl: Detecting Dependency Bugs to Prevent Compatibility Failures

Applications depend on libraries to avoid reinventing the wheel. Libraries may have incompatible changes during evolving. As a result, applications will suffer from compatibility failures. There has been much research on addressing detecting incompatible changes in libraries, or helping applications co-evolve with the libraries. The existing solution helps the latest application version work well against the latest library version as an afterthought. However, end users have already been suffering from the failures and have to wait for new versions. In this paper, we propose DepOwl, a practical tool helping users prevent compatibility failures. The key idea is to avoid using incompatible versions from the very beginning. We evaluated DepOwl on 38 known compatibility failures from StackOverflow, and DepOwl can prevent 32 of them. We also evaluated DepOwl using the software repository shipped with Ubuntu-19.10. DepOwl detected 77 unknown dependency bugs, which may lead to compatibility failures.

cs.SE↗

arXiv2020

OD-SGD: One-step Delay Stochastic Gradient Descent for Distributed Training

The training of modern deep learning neural network calls for large amounts of computation, which is often provided by GPUs or other specific accelerators. To scale out to achieve faster training speed, two update algorithms are mainly applied in the distributed training process, i.e. the Synchronous SGD algorithm (SSGD) and Asynchronous SGD algorithm (ASGD). SSGD obtains good convergence point while the training speed is slowed down by the synchronous barrier. ASGD has faster training speed but the convergence point is lower when compared to SSGD. To sufficiently utilize the advantages of SSGD and ASGD, we propose a novel technology named One-step Delay SGD (OD-SGD) to combine their strengths in the training process. Therefore, we can achieve similar convergence point and training speed as SSGD and ASGD separately. To the best of our knowledge, we make the first attempt to combine the features of SSGD and ASGD to improve distributed training performance. Each iteration of OD-SGD contains a global update in the parameter server node and local updates in the worker nodes, the local update is introduced to update and compensate the delayed local weights. We evaluate our proposed algorithm on MNIST, CIFAR-10 and ImageNet datasets. Experimental results show that OD-SGD can obtain similar or even slightly better accuracy than SSGD, while its training speed is much faster, which even exceeds the training speed of ASGD.

cs.LG↗

arXiv2015

New fast divide-and-conquer algorithms for the symmetric tridiagonal eigenvalue problem

In this paper, two accelerated divide-and-conquer algorithms are proposed for the symmetric tridiagonal eigenvalue problem, which cost $O(N^2r)$ {flops} in the worst case, where $N$ is the dimension of the matrix and $r$ is a modest number depending on the distribution of eigenvalues. Both of these algorithms use hierarchically semiseparable (HSS) matrices to approximate some intermediate eigenvector matrices which are Cauchy-like matrices and are off-diagonally low-rank. The difference of these two versions lies in using different HSS construction algorithms, one (denoted by {ADC1}) uses a structured low-rank approximation method and the other ({ADC2}) uses a randomized HSS construction algorithm. For the ADC2 algorithm, a method is proposed to estimate the off-diagonal rank. Numerous experiments have been done to show their stability and efficiency. These algorithms are implemented in parallel in a shared memory environment, and some parallel implementation details are included. Comparing the ADCs with highly optimized multithreaded libraries such as Intel MKL, we find that ADCs could be more than 6x times faster for some large matrices with few deflations.

math.NA↗

arXiv2015

On the Shoulders of Giants: Incremental Influence Maximization in Evolving Social Networks

Identifying the most influential individuals can provide invaluable help in developing and deploying effective viral marketing strategies. Previous studies mainly focus on designing efficient algorithms or heuristics to find top-K influential nodes on a given static social network. While, as a matter of fact, real-world social networks keep evolving over time and a recalculation upon the changed network inevitably leads to a long running time, significantly affecting the efficiency. In this paper, we observe from real-world traces that the evolution of social network follows the preferential attachment rule and the influential nodes are mainly selected from high-degree nodes. Such observations shed light on the design of IncInf, an incremental approach that can efficiently locate the top-K influential individuals in evolving social networks based on previous information instead of calculation from scratch. In particular, IncInf quantitatively analyzes the influence spread changes of nodes by localizing the impact of topology evolution to only local regions, and a pruning strategy is further proposed to effectively narrow the search space into nodes experiencing major increases or with high degrees. We carried out extensive experiments on real-world dynamic social networks including Facebook, NetHEPT, and Flickr. Experimental results demonstrate that, compared with the state-of-the-art static heuristic, IncInf achieves as much as 21X speedup in execution time while maintaining matching performance in terms of influence spread.

cs.SI↗