Zaifeng Pan, Ajjkumar Patel, Zhengding Hu, Yipeng Shen, Yue Guan, Wan-Lu Li, Lianhui Qin, Yida Wang, Yufei Ding
Large language model (LLM) based agentic workflows have become a popular paradigm for coordinating multiple specialized agents to solve complex tasks. To improve serving efficiency, existing LLM systems employ prefix caching to reuse key-value (KV) tensors corresponding to agents' fixed prompts, thereby avoiding redundant computation across repeated invocations. However, current systems typically evict KV caches using a Least Recently Used (LRU) policy, which fails to anticipate future agent usage and often discards KV caches shortly before their reuse. This leads to frequent cache misses and substantial recomputation or swapping overhead. We present KVFlow, a workflow-aware KV cache management framework tailored for agentic workloads. KVFlow abstracts the agent execution schedule as an Agent Step Graph and assigns each agent a steps-to-execution value that estimates its temporal proximity to future activation. These values guide a fine-grained eviction policy at the KV node level, allowing KVFlow to preserve entries likely to be reused and efficiently manage shared prefixes in tree-structured caches. Moreover, KVFlow introduces a fully overlapped KV prefetching mechanism, which proactively loads required tensors from CPU to GPU in background threads for agents scheduled in the next step, thereby avoiding cache miss stalls during generation. Compared to SGLang with hierarchical radix cache, KVFlow achieves up to 1.83$\times$ speedup for single workflows with large prompts, and up to 2.19$\times$ speedup for scenarios with many concurrent workflows.
Zaifeng Pan, Yipeng Shen, Zhengding Hu, Zhuang Wang, Aninda Manocha, Zheng Wang, Zhongkai Yu, Yue Guan, Yufei Ding
LLM-based multi-agent simulations are increasingly adopted across application domains, but remain difficult to scale due to GPU memory pressure. Each agent maintains private GPU-resident states, including models, prefix caches, and adapters, which quickly exhaust device memory as the agent count grows. We identify two key properties of these workloads: sparse agent activation and an estimable agent invocation order. Based on an analysis of representative workload classes, we introduce invocation distance, a unified abstraction that estimates the relative order in which agents will issue future LLM requests. Leveraging this abstraction, we present ScaleSim, a memory-efficient LLM serving system for large-scale multi-agent simulations. ScaleSim enables proactive prefetching and priority-based eviction, supports diverse agent-specific memory through a modular interface, and achieves up to 1.74x speedup over SGLang on simulation benchmarks.
Feng Zhang, Zaifeng Pan, Yanliang Zhou, Jidong Zhai, Xipeng Shen, Onur Mutlu, Xiaoyong Du
Text analytics directly on compression (TADOC) has proven to be a promising technology for big data analytics. GPUs are extremely popular accelerators for data analytics systems. Unfortunately, no work so far shows how to utilize GPUs to accelerate TADOC. We describe G-TADOC, the first framework that provides GPU-based text analytics directly on compression, effectively enabling efficient text analytics on GPUs without decompressing the input data. G-TADOC solves three major challenges. First, TADOC involves a large amount of dependencies, which makes it difficult to exploit massive parallelism on a GPU. We develop a novel fine-grained thread-level workload scheduling strategy for GPU threads, which partitions heavily-dependent loads adaptively in a fine-grained manner. Second, in developing G-TADOC, thousands of GPU threads writing to the same result buffer leads to inconsistency while directly using locks and atomic operations lead to large synchronization overheads. We develop a memory pool with thread-safe data structures on GPUs to handle such difficulties. Third, maintaining the sequence information among words is essential for lossless compression. We design a sequence-support strategy, which maintains high GPU parallelism while ensuring sequence information. Our experimental evaluations show that G-TADOC provides 31.1x average speedup compared to state-of-the-art TADOC.
Yue Guan, Changming Yu, Shihan Fang, Weiming Hu, Zaifeng Pan, Zheng Wang, Zihan Liu, Yangjie Zhou, Yufei Ding, Minyi Guo, Jingwen Leng
Speculative decoding improves LLM inference by generating and verifying multiple tokens in parallel, but existing systems suffer from suboptimal performance due to a mismatch between dynamic speculation and static runtime assumptions. We present Yggdrasil, a co-designed system that enables latency-optimal speculative decoding through context-aware tree drafting and compiler-friendly execution. Yggdrasil introduces an equal-growth tree structure for static graph compatibility, a latency-aware optimization objective for draft selection, and stage-based scheduling to reduce overhead. Yggdrasil supports unmodified LLMs and achieves up to $3.98\times$ speedup over state-of-the-art baselines across multiple hardware setups.
Zhongkai Yu, Chenyang Zhou, Yichen Lin, Hejia Zhang, Haotian Ye, Junxia Cui, Zaifeng Pan, Jishen Zhao, Yufei Ding
While Large Language Models (LLMs) show significant potential in hardware engineering, current benchmarks suffer from saturation and limited task diversity, failing to reflect LLMs' performance in real industrial workflows. To address this gap, we propose a comprehensive benchmark for AI-aided chip design that rigorously evaluates LLMs across three critical tasks: Verilog generation, debugging, and reference model generation. Our benchmark features 44 realistic modules with complex hierarchical structures, 89 systematic debugging cases, and 132 reference model samples across Python, SystemC, and CXXRTL. Evaluation results reveal substantial performance gaps, with state-of-the-art Claude-4.5-opus achieving only 30.74\% on Verilog generation and 13.33\% on Python reference model generation, demonstrating significant challenges compared to existing saturated benchmarks where SOTA models achieve over 95\% pass rates. Additionally, to help enhance LLM reference model generation, we provide an automated toolbox for high-quality training data generation, facilitating future research in this underexplored domain. Our code is available at https://github.com/zhongkaiyu/ChipBench.git.
Zhengding Hu, Zaifeng Pan, Prabhleen Kaur, Vibha Murthy, Zhongkai Yu, Yue Guan, Zhen Wang, Steven Swanson, Yufei Ding
In this work, we identify and address the core challenges of agentic memory management in LLM serving, where large-scale storage, frequent updates, and multiple coexisting agents jointly introduce complex and high-cost approximate nearest neighbor (ANN) searching problems. We present Pancake, a multi-tier agentic memory system that unifies three key techniques: (i) multi-level index caching for single agents, (ii) coordinated index management across multiple agents, and (iii) collaborative GPU-CPU acceleration. Pancake exposes easy-to-use interface that can be integrated into memory-based agents like Mem-GPT, and is compatible with agentic frameworks such as LangChain and LlamaIndex. Experiments on realistic agent workloads show that Pancake substantially outperforms existing frameworks, achieving more than 4.29x end-to-end throughput improvement.
Zheng Wang, Anna Cai, Xinfeng Xie, Zaifeng Pan, Yue Guan, Weiwei Chu, Jie Wang, Shikai Li, Jianyu Huang, Chris Cai, Yuchen Hao, Yufei Ding
In this work, we present WLB-LLM, a workLoad-balanced 4D parallelism for large language model training. We first thoroughly analyze the workload imbalance issue in LLM training and identify two primary sources of imbalance at the pipeline parallelism and context parallelism levels. Then, to address the imbalance issue, at the pipeline parallelism level, WLB-LLM incorporates a workload-aware variable-length document packing method to balance the computation and communication workload across micro-batches. Additionally, at the context parallelism level, WLB-LLM introduces a novel fine-grained per-document sharding strategy, ensuring each worker within a context parallelism group has an identical workload. Comprehensive experiments under different model scales demonstrate that WLB-LLM significantly mitigates the workload imbalance during 4D parallelism LLM training and achieves an average speedup of 1.23x when applying WLB-LLM in our internal LLM training framework.
Zhengding Hu, Vibha Murthy, Zaifeng Pan, Wanlu Li, Xiaoyi Fang, Yufei Ding, Yuke Wang
This paper addresses emerging system-level challenges in heterogeneous retrieval-augmented generation (RAG) serving, where complex multi-stage workflows and diverse request patterns complicate efficient execution. We present HedraRAG, a runtime system built on a graph-based abstraction that exposes optimization opportunities across stage-level parallelism, intra-request similarity, and inter-request skewness. These opportunities are realized through dynamic graph transformations, such as node splitting, reordering, edge addition, and dependency rewiring, applied to wavefronts of subgraphs spanning concurrent requests. The resulting execution plans are mapped onto hybrid CPU-GPU pipelines to improve resource utilization and reduce latency. Evaluations across a wide range of RAG workflows demonstrate speedups exceeding 1.5x and reaching up to 5x over existing frameworks, showcasing the effectiveness of coordinated generation and retrieval in serving environments.