Search arXiv⌕ Search

arXiv subjects

Yufei Ding

Publications and source records attributed to Yufei Ding.

At least 19 recordsLinked to original sources
arXiv2026

Zepp: Accelerating Distributed MoE Serving under Relaxed Balance Constraints

As Mixture-of-Experts (MoE) models continue to scale, serving them increasingly relies on expert parallelism (EP) across a growing number of devices. Yet skewed expert workloads create imbalance across computation, communication, and memory, making load balancing a central optimization objective in distributed MoE serving. We observe that balance is not free: operations introduced to balance one dimension can themselves be expensive or imbalanced. This motivates us to rethink balance as a constraint rather than an optimization objective. We present Zepp, which directly optimizes the bottleneck communication in distributed MoE serving subject to simplified balance constraints on physical resources, i.e., GPUs and NICs. Zepp progressively optimizes inter-node communication across placement, routing, and execution. It first places expert replicas to reduce token communication under GPU constraints, then reshapes communication flows through split and merge primitives under NIC constraints, and finally partitions and schedules ex- pert computation to overlap the resulting communication. To adapt to dynamic workloads, Zepp jointly coordinates computation, token communication, and expert-weight movement at each iteration. Together, these designs allow Zepp to pursue the most efficient execution rather than a single-dimension balanced one. We implement Zepp and evaluate it against 7 state-of-the-art MoE serving systems, achieving up to 6.68$\times$ MoE layer speedup and a geometric mean speedup of 1.86$\times$ over the fastest competing baseline.

cs.DC↗
arXiv2026

RR-Evict: Fine-Grained Prefix Cache Eviction beyond LRU for Agentic LLM Serving

LLM-based agents execute long-horizon tasks through repeated model calls interleaved with tool execution and user interaction. As each call extends the history accumulated in previous turns, prefix caching avoids repeated prefill of the agent's entire context. However, the aggregate cache footprint grows with context length and concurrency, forcing serving systems to reclaim cached KV tensors. We identify recency synchronization: accesses to an agent's cached history refresh its cache nodes together, causing least recently used (LRU) eviction to concentrate on a few agents' private histories. When a fully evicted agent returns, it must recompute nearly its entire accumulated context, producing a large time-to-first-token (TTFT) outlier even if most other requests retain substantial cache reuse. We present RR-EVICT, a fine-grained prefix-cache eviction strategy that distributes reclamation across idle agent trajectories. RR-EVICT visits agents in round-robin order and evicts a tail chunk from each, preserving reusable prefixes for more agents when capacity remains for idle state. Returning agents can reuse these partial histories and recompute smaller missing suffixes. The policy requires no prediction of future arrivals or tool latency. We implement RR-EVICT in SGLang and evaluate conversational and coding-agent workloads under colocated and prefill-decode-disaggregated serving. Compared with LRU, RR-EVICT reduces P99 TTFT by up to 75.4% and P99 uncached prompt tokens by up to 65.7%.

cs.DC↗
arXiv2026

Beyond Parallel Blindness: Information Floors and Model Gaps in Block Drafting

Block drafters propose several tokens in one forward pass, before earlier target tokens are realised. Their rejection mixes two losses: missing within-block path information and imperfect modelling of observable information. Accepted length cannot distinguish them. We separate the two with an information floor, the minimum expected rejection at a specified conditioning order; rejection above this floor is the model gap. Measurements across four domains and four open-weight targets, plus floor estimates on a frontier API target, yield three findings. First, the all-parallel floor reaches $0.286$ at the final slot on Qwen3-4B, limiting even the best proposal to $71\%$ per-slot acceptance. Second, one realised token removes $86$--$100\%$ of this floor, a locality also recovered by an independent mutual-information analysis. Third, current drafters remain far above their floors: the final-slot model gap accounts for $43$--$64\%$ of DFlash rejection and $85$--$92\%$ of DSpark's oracle-conditioned rejection. Guided by this separation of short-range conditioning from proposal quality, we replace a context-independent predecessor correction with a prefix-attention head that reads committed context conditional on the predecessor. With supporting backbone components, the resulting Qwen3-4B drafter improves mean serving accepted length by $3.93\%$ over the released DSpark checkpoint across nine tasks, without increasing the conditioning order. Code and data: https://github.com/tie-pilot-qxw/specfloor; drafter checkpoints: https://huggingface.co/TIE-Pilot/dspark-attnconv-block7-qwen3-4b.

cs.LG↗
arXiv2026

AgentReplay: Token-Wise Trace Replay Is Essential for Fair Serving System Performance Benchmarking

LLM-based agents execute multi-turn workflows with interleaved model inference and tool calls, making efficient serving increasingly important. However, evaluating serving optimizations is challenging because identical tasks can produce different execution trajectories. Changes in generated tokens can alter subsequent prompts, tool calls, and reasoning turns, making it difficult to distinguish system improvements from workload variation. Greedy decoding does not guarantee identical outputs, while replaying only sequence lengths loses token information that affects prefix caching and mixture-of-experts (MoE) routing. To address these problems, we propose AgentReplay, a configurable trace record-and-replay framework for agent serving. AgentReplay records input/output tokens, expert selections, request dependencies, and tool durations. During replay, it forces the recorded output tokens while performing normal autoregressive computation, with optional controls for MoE expert selection and tool delays. This allows different systems to execute the same recorded workload while retaining their own batching, scheduling, and parallelization decisions. We further separate trajectory generation from performance evaluation, enabling compatible smaller models to replay long-horizon traces collected with more capable models. Our experiments show that token-wise replay in AgentReplay effectively eliminates workload variation that greedy decoding and length-wise replay cannot avoid, enabling fairer performance comparisons across serving configurations.

cs.DC↗
arXiv2026

Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States

As language models become more capable, long-term collaboration in learning, reasoning, and decision-making calls for a deeper understanding of the people they serve. Yet training such human-aware language models faces a fundamental supervision gap because current datasets for LLM assistant training contain few if any well-informed responses explicitly grounded in users' unspoken beliefs and goals. Scaling such supervision is inherently constrained, as users' underlying states are not directly observable. We thus propose the Mind2Dialogue framework to mitigate this gap by simulating users' mental states and turning them into privileged supervision for human-aware training. Specifically, we first propose a psychology-guided simulator that preserves personal characteristics while updating mental states through interaction to generate coherent conversations. The key idea is to enforce a shared evolving mental state that drives user behavior and guides an Oracle assistant's responses. Our privileged distillation then trains models on the Oracle's well-informed responses to assist users without direct access to their mental states at deployment. Moreover, we propose to evaluate human-aware learning by combining personalization and theory of mind, examining how models understand people and act on that understanding. Training on the full Mind2Dialogue corpus improves every reported personalization metric over the corresponding Qwen, Llama, and OLMo instruction-tuned baselines, including gains of 26.6 to 40.9 percentage points in preference-following generation. The gains extend to belief and action reasoning on Qwen and Llama, beyond personalized assistance. Looking forward, Mind2Dialogue makes user simulation a foundation for genuine AI collaborators that understand beliefs and intentions behind people's words and support their long-term goals across education, work, and everyday life.

cs.CL↗
arXiv2026

Argus: Orchestrating Cross-Layer GPU Performance Measurements around Semantic Regions

GPU developers and automated optimizers need performance evidence for semantic code regions--such as neural-network operator implementations and pipeline stages--but this evidence is fragmented across profiling tools. Answering a region-level question can require manually constructing probes and program variants, isolating interfering measurements, and mapping evidence to regions and execution contexts. We present Argus, a region-centric measurement planner and runtime that automates this workflow. Clients identify regions with boundary markers and select signals and execution scopes. Argus preserves region identity across compilation, execution, and measurement variants, constructs interference-aware multi-run plans, and orchestrates transformations and profiling across backends. It joins compiler-, hardware-, and system-level evidence using region identity and dynamic execution context, producing reports that record measurement origins and attribution ambiguity. We evaluate Argus across agentic kernel optimization, persistent megakernel optimization, and cross-level PGO. Across 44 persistent-GEMM and attention configurations, Argus improves 39/44 cases and raises AlphaEvolve's geometric-mean speedup from 5.4% to 8.9%. On a persistent TinyLlama-1.1B decode megakernel, an optimization agent reaches 1.65 ms/token with Argus versus 4.92 ms/token without it, producing a kernel $2.1\times$ faster than PyTorch with CUDA Graphs. Finally, Argus-guided cross-level PGO improves compute--communication overlap, increasing throughput by 7% on average across five multi-GPU settings.

cs.DC↗
arXiv2026

LightStim: A Framework for QEC Protocol Evaluation and Prototyping with Automated DEM Construction

Fault-tolerant quantum computing increasingly demands rigorous, circuit-level evaluation of diverse quantum error correction (QEC) protocols and efficient prototyping of new ones. Such evaluation requires both the physical circuit and its Detector Error Model (DEM) to estimate end-to-end logical error rates. However, DEM construction today is performed by manual annotation, a tedious and error-prone process that effectively limits evaluation to simple memory experiments. We present LightStim, a framework that automates DEM construction concurrently with circuit compilation by maintaining a Pauli tableau augmented with measurement records, with no protocol-specific input required. We benchmark LightStim across protocols from memory experiments to end-to-end distillation circuits; cross-validation against public implementations confirms exact detector and observable counts and consistent logical error rates. Additionally, we demonstrate a novel heterogeneous cross-code lattice surgery design between surface and punctured quantum Reed-Muller codes. These capabilities together make LightStim a unified infrastructure for systematic QEC protocol evaluation and exploration. LightStim is open-sourced at https://github.com/QuTone/LightStim.

quant-ph↗
arXiv2026

PLoRA: An NDP-Enhanced Pooled-Memory System for Cost-Efficient Multi-LoRA Serving

Multi-LoRA serving is how one base model becomes thousands of specialized variants, one adapter per user, task, or agent, and the deployments can hold 1000-plus adapters. Serving them is hard because the workload inverts what GPUs provide: terabytes of memory against only tens of TFLOPS, and because every published system stages its adapters from CPU DRAM over PCIe, where each access pays a kernel stop and a host-run copy and capacity ends at the motherboard's DIMM slots. Meanwhile, memory-semantic fabrics such as CXL and NVLink are converging on pooled memory that an accelerator addresses with its own loads and stores, and near-data processing (NDP) can place compute beside the pooled data. How to serve multi-LoRA workloads on such hardware remains unexplored. This paper introduces PLoRA, an NDP-enhanced pooled-memory system for cost-efficient multi-LoRA serving. PLoRA keeps adapters and KV cache in the pool and returns only reduced results over the link, through a read-compute interface the GPU drives with its own loads and stores. Above this architecture, a GPU memory management system picks among four LoRA and two attention execution strategies for each adapter and caches the most performance-critical bytes in GPU memory, guided by a link-parameterized cost model. On one H100 serving 1000 adapters, PLoRA attains the lowest decode latency on every model and workload we measure, averaging 6.6x below a real-machine S-LoRA at under 3.4% added device area. The link itself stops mattering: throughput saturates at 32 GB/s on short contexts, a quarter of CXL 3.1, and the verdict survives scale: per-GPU demand falls from 7B to a modeled 1.2T deployment once adapter traffic shards with the tensor parallelism. The design runs unchanged from CXL-class to NVLink-class fabrics, and surplus bandwidth buys pooled capacity rather than speed.

cs.AR↗
arXiv2026

iSwitch: QEC on Demand via In-Situ Encoding of Bare Qubits for Ion Trap Architectures

Recent advances in quantum hardware and error correction have paved the way for early fault-tolerant (EFT) quantum computing. We propose iSwitch, a hybrid system architecture for trapped-ion quantum computers (TIQC) that exploits ultra-high-fidelity single-qubit gates and efficient logical CNOTs enabled by ion shuttling. iSwitch employs bare qubits for single-qubit operations and QEC-encoded logical qubits for two-qubit gates, avoiding full logical encoding, gate synthesis, and magic state distillation. To enable this selective encoding, we develop a low-noise conversion protocol between bare and logical qubits, a hybrid instruction set tailored to 2D TIQC layouts, and a compiler that minimizes conversion overhead and optimizes scheduling. Evaluations on variational quantum algorithm benchmarks show that iSwitch achieves comparable fidelity to conventional QEC methods, while reducing qubit and operation counts by roughly 33-50%, offering a practical, resource-efficient path toward EFT quantum computing on trapped-ion platforms.

quant-ph↗
arXiv2026

Syncopate: Efficient Multi-GPU AI Kernels via Automatic Chunk-Centric Compute-Communication Overlap

Communication has become a first-order bottleneck in large-scale GPU workloads, and existing distributed compilers address it mainly by overlapping whole compute and communication kernels at the stream level. This coarse granularity incurs extra kernel launches, forces device-wide synchronizations at kernel boundaries, and leaves substantial slack when the slowest tile or kernel stretches the communication tail. We present Syncopate, a compiler and runtime that enables automatic fine-grained overlap inside a single fused kernel. Syncopate introduces a communication chunk abstraction that decouples communication granularity from kernel structure and backend mechanisms, allowing chunk-level plans to be ported from existing distributed compilers, written directly by users, or instantiated from reusable templates. Given a local Triton kernel and a chunk schedule, Syncopate performs transformations to align computation with chunk availability. Implemented as a source-to-source compiler on Triton, Syncopate delivers an average end-to-end speedup of 1.3$\times$ and up to 4.7$\times$ on multi-GPU workloads. Our code is open-sourced at https://github.com/tie-pilot-qxw/syncopate.

cs.DC↗
arXiv2026

SmoothAgent: Efficient Long-Horizon LLM-Based Agent Serving with Lookahead Context Engineering

LLM-based agents execute multi-turn workflows with continuously growing contexts, where LLM calls are interleaved with tool invocations and environment feedback. To maintain model quality, modern agent frameworks rely on context engineering strategies such as offloading, reduction, and isolation to control the context length. However, these strategies introduce significant context transformation overhead: each transformation invalidates existing KV caches and triggers re-prefill, leading to increased time-to-first-token (TTFT). In this paper, we identify that context transformations are segment-decomposable, where the transformation of a prefix is independent of future tokens. This property enables transformations to be executed ahead of time. Based on this insight, we propose a lookahead programming model that allows agent frameworks to express context transformations as asynchronous operations without modifying their execution logic. The runtime proactively executes these transformations and prepares transformed KV caches in advance, enabling direct context replacement without blocking. We further design a lookahead-aware scheduler in LLM serving systems to support these asynchronous requests alongside latency-critical workloads with controlled interference. We implement our approach to support representative context engineering strategies and integrate it into existing agent frameworks and LLM serving systems. Experiments show that our approach effectively eliminates transformation overhead and reduces TTFT by up to 11.9x.

cs.DC↗
arXiv2026

Boron-assisted synthesis of compositionally complex amorphous oxides via short-range-order-constrained generative design

Engineering short-range atomic order in amorphous materials offers a promising yet still underexplored route to high-performance solids. Here, we establish a boron-assisted amorphization strategy through ApolloX, a theory-guided, short-range-order-constrained generative framework for identifying low-energy amorphous configurations in compositionally complex multimetal BOx systems. Using FeCoNiMoBOx as a representative model platform, ApolloX predicts an ensemble of candidate amorphous configurations across systematically varied boron contents. Ab initio molecular dynamics simulations based on these configurations show that increasing boron content suppresses atomic diffusion and disfavors crystallization, with the stabilization of BO3-centered local motifs emerging as a key structural feature associated with enhanced amorphization propensity. Guided by these predictions, we synthesize three representative FeCoNiMoBOx compositions with distinct boron contents and use synchrotron-based scattering and electron microscopy to verify compositional fidelity, structural homogeneity, and the targeted amorphous features, thereby experimentally validating the boron-regulated structural evolution predicted by theory. Beyond this representative system, the same strategy is extended to a broader library of multimetal BOx amorphous compositions spanning diverse metal combinations and boron loadings, demonstrating that the boron-assisted route is not limited to a single FeCoNiMoBOx family but is broadly transferable across compositionally complex amorphous oxides. Overall, our results establish boron incorporation as a practical design variable for tuning short-range order and amorphization in multicomponent oxides, and provide a general framework for the theory-guided discovery of compositionally complex amorphous materials with tunable properties.

cond-mat.mtrl-sci↗
arXiv2026

FlashCP: Load-Balanced Communication-Efficient Context Parallelism for LLM Training

Context parallelism (CP) is essential for training large-scale, long-context language models, as it partitions sequences to reduce memory overhead. However, existing CP methods suffer from workload imbalance, inefficient kernels, and redundant communication due to static sequence sharding and key-value (KV) tensor communication. We present FlashCP, a load-balanced and communication-efficient framework for CP training. FlashCP introduces a sharding-aware communication mechanism to eliminate redundant KV communication and proposes a novel Whole-Doc sharding strategy that maximizes communication savings while maintaining balanced workloads. To efficiently combine Whole-Doc and Per-Doc sharding, FlashCP further designs a heuristic algorithm to search for near-optimal sharding plans. Extensive experiments show that FlashCP achieves up to 1.63x speedup over state-of-the-art CP frameworks across diverse datasets.

cs.DC↗
arXiv2026

Transpiler-Architecture Co-Design to Curb Clifford Costs in Fault-Tolerant Quantum Computing

Quantum Error Correction (QEC) codes form the foundation of Fault-Tolerant Quantum Computing (FTQC) and predominantly use the Clifford+T gate set. Recently, Clifford operations have become the key performance bottleneck in implementing QEC. While state-of-the-art approaches like Pauli-Based Compilation (PBC) reduce Clifford overhead by transforming Clifford gates into Pauli measurements, they do so at the cost of gate-level parallelism, inflating circuit depth and execution times. To overcome these limitations, we introduce TACO, a Transpiler-Architecture Co-design framework that tackles the Clifford bottleneck through circuit and architectural optimization. TACO uses FTQC insights to guide hardware-aware Clifford gate elimination and circuit restructuring, and leverages the resulting optimized circuits to refine architectural design. TACO applies FTQC-specific transformations to aggressively reduce Clifford overhead from rotation synthesis and Toffoli decompositions, while preserving gate-level parallelism. The resulting architecture is optimized for the locality and data-movement patterns of these circuits, enabling high-throughput, resource-efficient execution. Our evaluation across diverse benchmarks shows that TACO achieves up to 21.9x (mean 4.4x) reduction in execution time compared to the state-of-the-art baseline.

quant-ph↗
arXiv2026

TLX: Hardware-Native, Evolvable MIMW GPU Compiler for Large-scale Production Environments

Modern GPUs increasingly rely on specialized hardware units and asynchronous coordination mechanisms, so performance depends on orchestrating data movement, tensor-core computation, and synchronization rather than exposing more thread-level parallelism. This creates a programming-model tension: if too much execution structure is hidden, the compiler must catch up to new hardware mechanisms; if too much is exposed, the burden of orchestration falls back onto the programmer. We present TLX (Triton Low-level Language Extensions), built around MIMW (Multi-Instruction, Multi-Warp), which expresses orchestration at warp-group granularity while preserving Triton's productive blocked programming model for regular computation. TLX realizes this idea as an embedded extension to Triton, exposing explicit interfaces for multi-warp execution, local-memory orchestration, asynchronous operations, and cluster-aware control. Our evaluation shows that TLX supports substantial customization with limited development effort while remaining competitive with state-of-the-art implementations. TLX-authored kernels have been deployed in large-scale training and inference production systems. Our code is open sourced at https://github.com/facebookexperimental/triton.

cs.AR↗
arXiv2026

ChipMATE: Multi-Agent Training via Reinforcement Learning for Enhanced RTL Generation

Existing API-based agentic systems for RTL code generation are fundamentally misaligned with industrial practice: they assume a golden testbench is available at generation time, rely on closed-source APIs incompatible with chip vendors' air-gapped security requirements, and cannot be trained on vendors' proprietary RTL codebases, leaving valuable internal data unused. Recent self-trained models address the deployment constraint but remain single-turn generators that overlook the critical role of verification in real industrial flows. To bridge these gaps, we present ChipMATE, the first self-trained multi-agent framework for RTL generation. Inspired by industrial practice where correctness emerges from cross-comparison between independently written RTL modules and reference models, ChipMATE pairs a Verilog agent with a Python reference-model agent that mutually verify each other's outputs without any golden oracle. We design a backtrack-based inference workflow to prevent error propagation across turns, and a two-stage training pipeline that first trains each agent individually to saturate its code-generation capability, then trains the team jointly to collaborate effectively. To support the training, we further build a hybrid data-generation framework that produces 64.4K high-quality reference model training samples. ChipMATE achieves 75.0\% and 80.1\% pass@1 on VerilogEval V2 with 4B and 9B base models, outperforming all existing self-trained models and even DeepSeek V4 with 1600B parameters. Our code and model weights are publicly available in https://github.com/zhongkaiyu/ChipMATE.

cs.MA↗
arXiv2026

Patterns behind Chaos: Forecasting Data Movement for Efficient Large-Scale MoE LLM Inference

Large-scale Mixture of Experts (MoE) Large Language Models (LLMs) have recently become the frontier open-weight models, achieving remarkable model capability similar to proprietary ones. But their random expert selection mechanism introduces significant data movement overhead that becomes the dominant bottleneck in multi-unit LLM serving systems. To understand the patterns underlying this data movement, we conduct comprehensive data-movement-centric profiling across four state-of-the-art large-scale MoE models released in 2025 (200B-1000B) using over 24,000 requests spanning diverse workloads. We perform systematic analysis from both temporal and spatial perspectives and distill six key insights to guide the design of diverse serving systems. We verify these insights on both future wafer-scale GPU architectures and existing GPU systems. On wafer-scale GPUs, lightweight architectural modifications guided by our insights yield a 6.6$\times$ average speedup across four 200B--1000B models. On existing GPU systems, our insights drive the design of a prefill-aware expert placement algorithm that achieves up to 1.25$\times$ speedup on MoE computation. Our work presents the first comprehensive data-centric analysis of large-scale MoE models together with a concrete design study applying the learned lessons. Our profiling traces are publicly available at \href{https://huggingface.co/datasets/core12345/MoE_expert_selection_trace}{\textcolor{blue}{https://huggingface.co/datasets/core12345/MoE\_expert\_selection\_trace}}.

cs.DC↗
arXiv2026

FlashEvolve: Accelerating Agent Self-Evolution with Asynchronous Stage Orchestration

LLM-based evolution has emerged as a promising way to improve agents by refining non-parametric artifacts, but its wall-clock cost remains a major bottleneck. We identify that this cost comes from synchronized stage execution and imbalance inside each LLM-heavy stage. We present FlashEvolve, an efficient framework that replaces synchronized execution with asynchronous workers and queues, allowing different stages and steps to overlap. To handle data staleness introduced by asynchrony, FlashEvolve tracks artifact versions and applies different policies to update, discard, or patch stale artifacts. Unlike weight-space staleness in asynchronous RL, language-space staleness is inspectable and repairable: a stale artifact is not just delayed work, but readable evidence that the LLM can reflect on, revise, and turn into useful evolution signal. FlashEvolve further improves throughput and token efficiency with speculative stage completion and adaptive workflow control. On GEPA workloads, FlashEvolve improves proposal throughput by $3.5\times$ on local vLLM and $4.9\times$ on API serving over synchronous GEPA. The same design also applies to ACE and Meta-Harness.

cs.LG↗