跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+1284新增论文
+1011其中 AI 相关
+150已有论文修订
+97新增 AI 导读

arXiv 最近完成扫描:2026-11-20 08:48 · 全站最近入库:2026-11-20 10:50各信源进度 →

统计口径与数据新鲜度

统计 2026-11-19 12:31 — 2026-11-20 12:31(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-20 08:48最近发现新增版本:2026-11-20 10:50库内论文最新更新时间:2026-11-19 12:39

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

28,155 已收录论文23,386 AI 相关8,624 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 262 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-20 08:46(北京时间)
版本更新 · v2arXiv:2610.02665

大规模语言连续扩散模型

Large Language Continuous Diffusion Models

Zhihan Yang, Wei Guo, Jean-Marie Lemercier, Simon Welker, Mohammad Mahdi Kamani, Yonggan Fu, Sajad Norouzi, Julius Berner 等 17 位作者

AI 速读 基于摘要

提出首个大规模(3B/8B)连续扩散语言模型 Nemotron-Labs-Sigma,通过可操控的低维 ODE/SDE 潜轨迹实现高效推理与加速。

cs.CL · cs.AI · cs.LG

查看原文摘要

Despite the success of discrete diffusion language models (dLMs) for fast parallel decoding, their non-smooth, high-dimensional space hinders trajectory steering for reasoning and inference acceleration. To overcome this, we present Nemotron-Labs-Sigma, the first family of large-scale (3B/8B) continuous dLMs built on steerable, low-dimensional ODE/SDE latent trajectories. Trained blockwise via likelihood optimization, Nemotron-Labs-Sigma jointly denoises Gaussian-corrupted token embeddings while learning an optimal embedding geometry, and leverages pretrained autoregressive (AR) weights for warm-starting. At inference, we identify classifier-free guidance and score temperature as essential for high-fidelity reasoning and coding. Against state-of-the-art discrete counterparts (masked dLMs and AR baselines), Nemotron-Labs-Sigma outperforms them on standard benchmarks after pretraining (e.g., GSM8K, Minerva, HumanEval, MBPP) and on challenging reasoning tasks after supervised fine-tuning (e.g., MATH-500, AIME), with the largest consistent gains on coding. Beyond accuracy, we exploit two structural properties unique to continuous dLMs. First, smooth trajectories degrade gracefully at low NFEs and enable efficient few-step distillation: distribution matching proves strongest among the methods we study, giving roughly 8x acceleration on math and 4x on coding (8 and 4 tokens per forward pass) with negligible accuracy loss and preserved diversity. In addition, the one-step speculative decoding on the (pre-SFT) distilled 8B model yields a further ~3.5x throughput gain over the AR counterpart (on GB200), outperforming masked dLM baselines. Second, embedding-space steering controls the quality--diversity trade-off, yielding strong pass@$k$ that keeps improving as $k$ grows. Together, these results establish continuous dLMs as a promising paradigm for efficient language generation.

新论文 · v1arXiv:2611.17212

稀疏卷积的理论与实证分析

Theoretical and Empirical Analysis of Sparse Convolution

Keshvi Tuteja, Yu-Hsiang Tsai, Gregor Olenik, Markus Götz, Achim Streit, Hartwig Anzt, Charlotte Debus

AI 速读 基于摘要

该工作通过roofline分析刻画稀疏权重标准卷积的性能特征,并提出高稀疏度下的高效稀疏卷积技术,初步在NVIDIA A100与Jetson Nano上相对稠密卷积取得加速。

cs.PF · cs.DC · cs.LG

查看原文摘要

Deep neural networks for image processing are increasingly deployed on resource-constrained edge devices such as drones, medical systems, and smartphones, where limited computational resources and memory bandwidth pose significant challenges for efficient inference. Pruning during training often produces highly sparse convolutional kernels, offering the potential to significantly reduce computation and memory requirements. However, exploiting this sparsity efficiently at runtime remains challenging due to irregular memory access patterns and workload imbalance across processing elements. In this work, we investigate the standard convolution with sparse weights and analyze its performance characteristics through roofline analysis, showing that a 3 x 3 sparse convolution achieves an arithmetic intensity of approximately 10 FLOP/byte compared to roughly 96 FLOP/byte for its dense counterpart. We propose techniques for an efficient sparse convolution at high sparsities. Preliminary results demonstrate that our approach achieves upto 1.65x speedup as compared to dense convolution on NVIDIA A100 and upto 1.52x on NVIDIA Jetson Nano (Tegra X1) , motivating further exploration of direct convolution as a practical path to efficient sparse inference on modern accelerators/edge devices.

新论文 · v1arXiv:2611.17152

AeroZoo:面向空气动力学基础模型的良定流动问题基准

AeroZoo: A Benchmark of Well-Posed Flow Problems for Foundation Models of Aerodynamics

Peter Sharpe, Carmelo Gonzales, Corey Adams, Mehdi Ataei, Sanjay Choudhry

AI 速读 基于摘要

作者提出 AeroZoo 基准与训练语料,用 3,815 个 RANS 时均解覆盖多几何族、多外罩配置与宽雷诺数范围,并设计按几何族、外罩、边界条件布局和雷诺数范围留出的冻结划分,以检验空气动力学基础模型的跨分布泛化能力。

cs.LG · cs.CE · physics.flu-dyn

查看原文摘要

Neural surrogates for computational fluid dynamics (CFD) are trained and ranked on datasets that each vary one design family in one flow domain. Such datasets cannot test the least that an aerodynamic foundation model must do: predict CFD-quality incompressible flow in seconds for new shapes in new enclosures. On DrivAerML, larger models and longer training barely lower Transolver's error; doubling the training cases cuts it by about 12% up to 400 cases, exhausting the dataset. We introduce AeroZoo, a benchmark and training corpus of 3,815 time-mean solutions to the unsteady Reynolds-averaged Navier--Stokes (RANS) equations with a transition model, across seven geometry families, eight enclosure configurations and Reynolds numbers from $10^3$ to $10^8$. Each sample poses a well-posed PDE problem in the form a solver receives it: the boundary meshes grouped by condition type, the prescribed values and the fluid properties. The 2,816 samples with a stable time mean form the canonical benchmark: two tasks, engineering metrics and 22 frozen splits, disjoint in geometry, that hold out geometry families, enclosures, boundary-condition layouts and Reynolds ranges. On the random split, boundary-condition tokens cut GeoTransolver's median velocity error from 76.9% to 8.4% of the inlet speed in single-seed runs (uniform inlet flow: 11.0%). ISLA, a transformer for boundary-driven PDEs, reaches 5.3% over four seeds, about 4 times the change from a finer mesh. On the same test samples, a held-out geometry family barely changes ISLA's error relative to uniform flow, and a held-out enclosure or boundary-condition layout raises it 2.4 to 9.2 times. The data and the scoring code will be released.

新论文 · v1arXiv:2611.16856

可持续的LoRA联邦微调:环境影响与性能足迹

Sustainable Federated Fine-Tuning with LoRA: Environmental Impact and Performance Footprint

Pedro Augusto Londe Ribeiro, Rodrigo Moreira, Larissa Ferreira Rodrigues Moreira, Joberto S. B. Martins, Tereza C. Carvalho, Flávio de Oliveira Silva

AI 速读 基于摘要

作者报告在三种GPU上评估联邦LoRA微调的能耗与碳排放,发现本地步数主导能耗变化,较少步数可在保持精度的同时大幅降低能耗。

cs.LG · cs.NI

查看原文摘要

Federated Learning (FL) with Parameter-Efficient Fine-Tuning (PEFT) enables efficient adaptation of Large Language Models (LLMs), yet its energy and carbon costs across hardware remain poorly understood. We evaluated federated Low-Rank Adaptation (LoRA) fine-tuning on NVIDIA A40, RTX 6000 Ada, and A100 GPUs using a $2^2$ full factorial design with 135M/360M models and 10/100 local steps. CodeCarbon tracked energy and estimated CO$_2$-equivalent emissions. Descriptive allocation of variation assigned 84.8-91.2% of energy and emissions variation to local steps within these ranges. On A40 and RTX 6000 Ada, configurations with fewer steps achieved comparable accuracy with roughly 6-8-fold lower energy consumption and higher accuracy per kWh. These findings support smaller local budgets for energy-efficient federated fine-tuning within the evaluated setting.

版本更新 · v2arXiv:2611.09195

InstantHMR : Real-Time Human Mesh Recovery on Your Phone

Mohamed Adjel, Kahina Chalabi, Vincent Bonnet

原文摘要

Real-time 3D human mesh recovery from a single image on mobile devices is a challenging problem with many applications, including robotics, biomechanics, AR/VR, and exergames. Current state-of-the-art models are either too slow to run efficiently on mobile devices or sacrifice accuracy for speed. We present InstantHMR, a family of open-source single-image regressors suited for mobile inference, based on the Momentum Human Rig (MHR) parametric model and trained exclusively on publicly available annotations from the SAM 3D Body dataset. Our approach achieves remarkable performance on a Samsung Galaxy S23 NPU, recovering a 6,890-vertex body mesh in 5.5 ms per person, and 0.98 ms on an NVIDIA RTX 4070 GPU. This speed surpasses all existing methods we evaluated. On mobile, it is 12x faster than the single mobile-compatible method that slightly outperforms ours in accuracy, while achieving competitive accuracy with 40.0 mm PA-MPJPE on 3DPW, 48.8 mm on EMDB and 41.7 mm on RICH. To ensure fair comparisons, we re-evaluate 11 state-of-the-art methods under a unified protocol aligned with the closest MHR-based competitors, including Meta's official MHR-to-SMPL conversion, and report their reproduction results. Additionally, we measure latency across eight different targets, ranging from desktop TensorRT to two phone NPUs. Code and models are available at https://github.com/Naturalpad-studio/InstantHMR.

cs.CV

版本更新 · v2arXiv:2611.12304

Evolving Infra: Agent-Adaptive LLM Serving at Scale

Hao Kang, Weili Xu, Ziyang Li, Zheng Du, Hanchen Li, Tian Xia, Song Bian, Tushar Krishna 等 11 位作者

原文摘要

Agentic workloads are a major source of LLM inference demand. Existing prefill--decode (PD) disaggregated serving systems offer limited flexibility for runtime adaptation of compute allocation and KV cache management, and rely on coarse signals that only partly capture the diverse and dynamic agent workload. These limitations lead to load imbalance between prefill and decode nodes and eviction of KV caches that agents will reuse. We present Evolving Infra, an agent-adaptive serving framework that adapts compute allocation to agent workloads and KV cache management to agent context management at runtime. We decouple compute allocation from key--value (KV) cache management, allowing nodes to switch efficiently between prefill and decode. Building on this design, an online planner matches the PD split to workload composition, while an agent-aware distributed KV manager guides cache eviction and transfer with signals from agent context management. Across coding, math, and computer-use tasks with GLM-5.2 and Kimi K2.6 on up to 256 GB300 GPUs, Evolving Infra reduces mean end-to-end request latency and P95 TTFT by up to 4.1x and 14x over Dynamo, NVIDIA's production framework. We open-source the system implementation of Evolving Infra at https://github.com/mit-han-lab/EvolvingInfra.

cs.DC

新论文 · v1arXiv:2611.16363

Vector-Space Projection and Unified Complex-Valued Acceleration for Scaling Matrix-Based Brain-Inspired Representations

William Youngwoo Chung, Hyunwoo Oh, Calvin Yeung, Hansen Jin Lillemark, Hamza Errahmouni Barkam, Mohsen Imani

原文摘要

Symbolic reasoning at scale requires representations that are both expressive and hardware-efficient. Hyperdimensional Computing (HDC) provides a promising brain-inspired framework, encoding concepts as high-dimensional vectors composed through efficient algebraic operations. However, the widely used Fourier Holographic Reduced Representation (FHRR) relies on commutative binding, which cannot encode order or directionality. Prior FPGA/ASIC accelerators recover this through costly permutation logic, incurring irregular memory access and high control overhead. Generalized Holographic Reduced Representations (GHRR) instead redefine binding as non-commutative matrix multiplication, naturally encoding sequences, graphs, and hierarchies without permutations-but at the cost of turning every element into a matrix, introducing compute and storage overhead that limits scalability. In this work, we propose a flattening methodology that preserves GHRR's matrix-based encoding while executing training and inference directly in vector space, functionally equivalent to FHRR inference yet free of permutation logic. We further design a custom 28 nm ASIC that fuses binding and similarity into a unified complex-valued datapath, with dual-DMA streaming and runtime normalization for accurate inference. Against a PyTorch baseline on an NVIDIA RTX 4090 GPU, the prototype delivers 1.36x-1.56x higher throughput and 16.2x-18.6x better energy efficiency.

cs.AR

版本更新 · v3arXiv:2508.02001

Versatile yet Efficient Network Traffic Analysis: Offloading Network Foundation Model to SmartNIC

Chungang Lin, Xuying Meng, Tianyu Zuo, Weiyao Zhang, Meng Shen, Ruijie Zhao, Guanming Che, Ruiqi Meng 等 12 位作者

原文摘要

Pervasive encryption makes large-scale labeling infeasible for traffic analysis, while security operations demand edge analysis to avert service degradation and vulnerabilities. These pressures have produced two research lines that advance 1) versatile analysis, via network foundation models for low label dependency, and 2) efficient analysis, via hardware offloading for low analysis latency. However, co-achieving versatility and efficiency via such lines has appeared incompatible, with prior work advancing one while compromising the other, yet we show that this incompatibility is a consequence of polarized design choices across the three components of traffic analysis systems, i.e., traffic processing, model architecture, and analysis execution. In response, we present Nepco, a versatile yet efficient network traffic analysis system that offloads network foundation models to SmartNIC. Our key observation is that discriminative traffic information often exhibits localized patterns within byte sequences, motivating versatile yet efficient localized byte-sequence modeling. To exploit this without incurring the latency bottlenecks of complex encoding steps, we employ a hardware-friendly processing strategy that directly embeds raw byte sequences. To maintain versatility across diverse tasks characterized by spatially shifting features, we explore an observation-driven convolution architecture equipped with hardware-friendly scoring and gating designs. By exploiting translation invariance, this design dynamically locates and extracts salient semantic signatures irrespective of their byte-level offsets. We prototype Nepco on the NVIDIA BlueField-3 SmartNIC with multi-engine collaborative analysis execution. The experimental results demonstrate that Nepco achieves macro F1 competitive with the best performances achieved by 8 state-of-the-art network foundation models...

cs.NI · cs.LG

版本更新 · v4arXiv:2606.03159

NVIDIA OmniDreams: Real-Time Generative World Model for Closed-Loop Autonomous Vehicle Simulation

Aarti Basant, Amlan Kar, Despoina Paschalidou, Fangyin Wei, Francesco Ferroni, Guillermo Garcia Cobo, Haithem Turki, Huan Ling 等 33 位作者

原文摘要

As autonomous vehicle capabilities advance, the safe evaluation of driving policies in long-tail scenarios remains a critical bottleneck. In closed-loop simulation, the driving policy model actively interacts with the environment, where its actions dynamically update the simulator state and directly influence the next set of generated sensor observations. While recent reconstruction-based neural simulators offer photorealism, they are fundamentally constrained by their initial captured data and struggle to generalize to highly dynamic or novel scenes. To overcome these limitations, we introduce OmniDreams, a foundation generative world model mid- and post-trained from the Cosmos diffusion model to autoregressively generate action-conditioned videos in real time. By leveraging the rich visual priors of Cosmos and mid- and post-training on 21k hours of driving scenarios, OmniDreams synthesizes complex, unobserved phenomena that are hard for traditional simulators to capture, such as extreme weather and unpredictable dynamic agent behaviors. Crucially, it autoregressively conditions its photorealistic sensor generation on past frames, the current simulator state, and immediate driving actions. Deployed in a closed-loop system with the Alpamayo 1 policy model and AlpaSim orchestrator, OmniDreams acts as a highly responsive, reactive environment, providing a scalable and comprehensive solution for training and evaluating next-generation autonomous driving policies. We additionally show preliminary results indicating that a world-action model (WAM) post-trained from OmniDreams achieves strong performance on the Physical AI Autonomous Vehicles NuRec dataset, surpassing the VLA-based Alpamayo 1.5 research policy model while using only 1/5 the total parameters. These results highlight the potential for a real-time world model like OmniDreams to also serve as a backbone for policy architectures.

cs.CV · cs.AI · cs.RO

新论文 · v1arXiv:2611.16205

QVS-LIVO RiskMap: Health- and Information-Aware View Selection for Multi-Camera LIVO and Off-Road Risk Mapping

Zihan Pan, Jinghan Wang, Hang Li, Feng Cheng, Tianchen Liu, Gaoliang Peng

原文摘要

Multi-camera observations in off-road environments vary significantly in reliability, posing challenges when complementing degraded geometric constraints. Motivated by this, we propose Quality-aware View Selection Lidar-Vision Odometry Risk Mapping (QVS-LIVO RiskMap), which decouples visual processing for state estimation from that for semantic mapping. The front end evaluates camera observations through read-only pretracking on a shared lidar-inertial posterior, combining observation health with geometric complementarity. Channel health memory and switching hysteresis govern selection of at most one eligible view for direct photometric correction. The posterior is retained when no view is eligible or the correction fails acceptance. The mapping back end integrates semantic observations from all available views in a shared world grid. Vehicle-scale geometry constrains semantic risk corrections, while spatial and temporal evidence supports risk confirmation. The system also produces a local bird's-eye-view (BEV) risk map and a global 2.5D semantic-terrain risk map with obstacle instance voxels. On two field sequences, QVS-LIVO achieved horizontal root mean square errors of 0.910 m and 1.664 m. It reduced error by 43.3% relative to FAST-LIO2 on the degraded long-route sequence. Across 12 controlled camera-fault configurations involving dropout, black frames and 60 ms timestamp shifts, it maintained complete trajectories with errors of 1.54-2.23 m. When deployed onboard an NVIDIA Jetson AGX Orin, the system generates local and global maps at 9.94 Hz and 2.00 Hz, respectively, with a median local mapping latency of 32.4 ms. These runtime results demonstrate the real-time feasibility of selective visual state updates alongside multi-view semantic mapping under challenging off-road conditions. The code and related resources are available at: https://github.com/zihanPan-HIT/QVS-LIVO.

cs.RO

版本更新 · v2arXiv:2610.13679

A Simplified Layout Algebra for GPU Kernels

Nan An, Xiaotian Zhou

原文摘要

The layout algebra behind NVIDIA's CuTe library describes how tensor tiles map onto threads and memory. A layout is a shape with a stride for each entry, so a coordinate's address is the sum of its indices times their strides. As formalized by Shah and by Cecka, its central operations (e.g., composition, complement, division, product) are guarded by at least six interacting admissibility conditions. We show that this complexity is an artifact of a representational choice, not of the domain. We present a layout algebra in which every layout names the space its inputs and outputs live in: logical, physical, or thread-value indices. A layout into logical or thread-value space must be a bijection from its $N$ coordinates onto $[0,N)$. A layout into physical space must be alias-free: distinct coordinates land on distinct addresses, except under broadcast, where many threads read one value on purpose and the layout must mark where. Nothing maps out of a physical index. Composing $f$ with $g$ requires only that $f$'s range be exactly $g$'s flattened coordinate set. The index arithmetic can always be emitted naively: compute $f$, split the result by division and remainder into $g$'s coordinates, and apply $g$. Often the composite simplifies to a layout; we show that when it does, there is a coarsest such form, whose digits every other one refines, and we emit that. A warp's or thread's slice is a \emph{restriction}. A restricted layout is no longer composable, which forces the user to settle the whole view of a tile before taking slices. We validate the algebra by reproducing within it CuTe's operations, such as composition, complement, and division. We implement it in OCaml and use the implementation to rederive CuTe's atoms and usage patterns. The decision procedure is proved correct in Rocq.

cs.PL

版本更新 · v2arXiv:2608.07408

Addressable Memory for Video World Models

Xindi Wu, Sven Elflein, James Lucas, Olga Russakovsky, Laura Leal-Taixé, Despoina Paschalidou, Jonathan Lorraine, Aljoša Ošep

原文摘要

We study visual persistence in interactive video world models. These models rely on a Key-Value (KV) cache as a growing visual memory to carry forward previously generated frames. However, we find that models can no longer reliably address stored content once rollouts extend beyond the training horizon, because temporal Rotary Positional Embedding (RoPE) offsets can exceed the training range and the model struggles to retrieve the relevant visual information through attention. Moreover, naively compressing the cache in the RoPE-rotated space degrades memory by averaging keys with incompatible positional phases. To address this, we propose WorldTrace, a training-free memory framework for long-horizon visual persistence. WorldTrace keeps compressed memory addressable by assigning each summary slot a distinct, in-distribution virtual position. Within this cache, WorldTrace uses the model's own attention to decide which observations to retain verbatim and which redundant ones to merge into compressed summaries, all within one bounded addressable memory bank. We evaluate this ability with LoopBench, a benchmark that tests whether a compressed cache can reconstruct a previously visited scene after a long detour. On LoopBench, WorldTrace improves average DINOv2 recall similarity by 32.6% over a sliding window at the same cache budget, extending visually persistent generation without retraining.

cs.CV · cs.LG

新论文 · v1arXiv:2611.15926

NVIDIA Vera: A Microarchitecture-Level Comparison with Grace, Turin, and Granite Rapids

Joseph Moore, Thomas Green, Simon McIntosh-Smith

原文摘要

The majority of new HPC systems are centered around NVIDIA, AMD, or Intel CPUs. Vera is NVIDIA's latest data-center CPU, including its first custom data-center core design, Olympus. This paper investigates Vera in comparison to AMD Turin and Intel Granite Rapids, using NVIDIA Grace as a baseline. We navigate up the performance ladder, from microbenchmarks that isolate performance metrics, through NAS kernels and miniapps, to single-socket performance of a range of scientific codes. We find that Vera achieves the best performance of the architectures evaluated in most codes tested, and that this gain is primarily attributed to its high memory bandwidth, effective Spatial Multithreading implementation, and low-latency memory subsystem aided by the architecture's predictive structures. While Vera represents a major leap over its predecessor, we show that its comparatively modest floating-point performance trails Intel and AMD in a small number of compute-bound workloads reflecting NVIDIA's GPU focus for such workloads.

cs.DC · cs.AR · cs.PF

版本更新 · v2arXiv:2611.12737

Inference Validation for Regulated AI: A Governed Decision Pipeline for LLM Serving

Dev Akre, Roman Bobalo, Manoj Dobbali

原文摘要

Deploying large language models in regulated industries requires more than picking the fastest inference engine: it requires evidence that a configuration meets contractual and legal obligations before it serves a single customer request. We describe Runway, an internal validation harness built at MasterControl that builds a governed decision pipeline around NVIDIA AIPerf. The key insight is separating concerns: AIPerf provides the trusted measurement, while Runway provides the experiment design that comes before it and the compliance, cost, and audit decisions that come after. We describe the architecture and the storage model that makes caching and audit one mechanism. We report three findings from Runway sweeps: TensorRT-LLM and vLLM rankings invert with concurrency; a newer, pricier GPU instance can be the cheaper one per request; and speculative decoding behaves differently across engines under load, so engine configuration can outweigh engine choice.

cs.SE · cs.LG · cs.PF

版本更新 · v5arXiv:2512.02551

CUDA-L2: Surpassing cuBLAS Performance for Matrix Multiplication through Reinforcement Learning

Songqiao Su, Xiaoya Li, Albert Wang, Guoyin Wang, Jiwei Li, Chris Shum

原文摘要

In this paper, we propose CUDA-L2, a system that combines large language models (LLMs) and reinforcement learning (RL) to automatically optimize Half-precision General Matrix Multiply (HGEMM) CUDA kernels. Using CUDA execution speed as the RL reward, CUDA-L2 automatically optimizes HGEMM kernels across 1,000 configurations. CUDA-L2 systematically outperforms major matmul baselines to date, from the widely-used torch.matmul to state-of-the-art Nvidia's closed-source libraries, i.e., cuBLAS, cuBLASLt. In offline mode, where kernels are executed consecutively without time intervals, CUDA-L2 yields +22.0% over torch.matmul on average; +19.2% over cuBLAS using the optimal layout configuration (normal-normal NN and transposed-normal TN); +16.8% over cuBLASLt-heuristic, which queries cuBLASLt library and selects the algorithm based on the heuristic's suggestion; and +11.4% over the most competitive cuBLASLt-AutoTuning model, which selects the fastest algorithm from up to 100 candidates from cuBLASLt's suggestions. In server mode, where kernels are executed at random intervals simulating real-time inference, the speedups further increase to +28.7%, +26.0%, +22.4%, and +15.9% for torch.matmul, cuBLAS, cuBLASLt-heuristic, and cuBLASLt-AutoTuning respectively. CUDA-L2 shows that even the most performance-critical, heavily-optimized kernels like HGEMM can be improved through LLM-guided RL automation by systematically exploring configuration spaces at scales impractical for humans. Project and code can be found at github.com/ornith-ai/CUDA-L2

cs.LG · cs.AI

新论文 · v1arXiv:2611.15671

The Noise Floor of a Quantization Comparison: Engine, restart, batch and kernel drift, measured against 8-bit and 4-bit effects in two decision models

Aleks Blagojevich, Trevor Sunderland

原文摘要

A quantized model is usually judged by comparing it with its BF16 original, but that comparison changes more than the weights. It often changes the serving engine, it always runs on a different server instance, it may change how requests are batched, and on some stacks the library selects a different kernel depending only on which packages import. We measure each of these sources on the same 2,400 public decisions for two of Cloudflare's decision models, Clef-flash (9B) and Clef (27B), on an NVIDIA GB10, and state every quantization effect as the excess over them. Sources that are not quantization move between 0 and 0.92% of decisions. llama.cpp restarts are bit-identical on every record for both models. vLLM restarts flip 0, 5 and 20 decisions at BF16, FP8 and NVFP4 on Clef-flash and 4, 0 and 22 on Clef; a third instance of Clef's quantized builds (exploratory) leaves FP8 at 0 and puts NVFP4 at 4 to 22, depending on which two instances are compared. An engine change at BF16 flips 6 to 16, a batch-size change 8 to 15, and the Gated DeltaNet kernel transformers selects unasked 10 to 14. Quantization inside one engine moves more: for Clef-flash and Clef, 0.88% and 1.04% at Q8_0, 2.12% and 1.62% at FP8, 5.79% and 3.50% at Q4_K_M, and 5.79% and 5.67% at NVFP4. On both models, every 4-bit interval excludes every non-quantization source. Two of our five pre-registered predictions failed, and the same two fail when the tests are rerun on the 27B: FP8 is not inside restart noise, and changing engine moves fewer decisions than 8-bit quantization does. Accuracy hides most of this. Over the full 31,275 public questions, Clef's Q4_K_M build moves accuracy by a net six questions (-0.02 points [-0.20, +0.16]) while 436 of its decisions get worse (exploratory), and NVFP4 costs Clef-flash 0.95 points and Clef 0.65. We release the code, quantization recipes, per-record outputs and the pre-registration.

cs.LG

新论文 · v1arXiv:2611.15658

Embedded Sim-to-Real Visual Servoing: Learned Reaching and Force-Guided Placement

Dezhi Sun, Duy Khang Banh, Kenan Niu

原文摘要

Embedded perception-to-control systems must combine visual targeting with reliable contact handling under tight computational constraints. We present a shared architecture on an NVIDIA Jetson AGX Orin controlling a Franka Research 3 through Cartesian impedance control. RealSense D435i observations, YOLO11-seg segmentation, depth back-projection, and hand--eye calibration provide metric robot-frame targets. We compare two strategies: compact proximal policy optimization (PPO) policies transferred from simulation for object- and hole-region reaching, and force-guided Cartesian placement using approximately 2 N normal contact and spiral search. The PPO policies achieved 87.2% and 83.1% success over 1000 simulation episodes each. Across 50 real-robot trials per strategy, the learned hybrid and force-guided analytical pipelines achieved 70.0% and 74.0% completion, respectively. Mean perception latency with TensorRT was 34.55 ms. These results support an embedded architecture combining learned free-space reaching with explicit contact correction.

cs.RO

版本更新 · v2arXiv:2611.07535

Characterizing CUDA Locality Domains on Blackwell GPUs

Rocky Wang, Joyce Zhou, Keith Jiang, Jason Zhao

原文摘要

Low latency is critical to user experience. As NVIDIA Blackwell and Rubin GPUs advance multi-die scaling, data placement and cross-die communication become increasingly important. However, the latency implications and underlying costs of cross-die communication and die-aware execution remain largely unexplored. In this report, we present a comprehensive evaluation of the Locality Domain (LD) feature introduced in CUDA 13.4. Conducted on a single node equipped with eight NVIDIA B300 GPUs, our study examines three aspects: memory management and access, computation, and communication, to identify the sources of latency and, more importantly, explore how LD can mitigate them. Our measurements show that the benefits of LD depend on data access patterns. Through die-aware adaptations and integration with widely used libraries and tools, we achieve speedups of up to 43.5% for NCCL 2.32.3 (AllGather), 111.3% for nvbandwidth (DtoD), and 37% for DeepGEMM (EP). We also identify non-negligible LD-related GPU memory allocation overhead when allocation sizes exceed 32 MB due to barrier and TLB invalidation, we propose a fix to the GPU driver kernel module (submitted to NVIDIA) and show that overhead could be reduced 7x ~ 10x without requiring any user-code changes.

cs.DC

新论文 · v1arXiv:2611.15372

LLMRunner: Closed Workload Model Performance Testing Framework for LLM Servings

Arush Upadhyaya, Debadatta Mishra

原文摘要

Performance testing large language model (LLM) serving engines is a crucial requirement considering the criticality of the infrastructure for deploying conversational AI at scale. Most of the benchmarking tools used to carryout performance evaluation of LLM serving do not capture realistic workload orchestration unlike closed workload performance testing infrastructure used in other domains such as web and database engines. For example, state-of-the-art LLM benchmarks for performance testing model traffic as streams of independent single-turn and stand-alone requests omitting the end-user behavior such as think-time between conversational turns. We present LLMRunner, a performance tester for LLM serving that addresses this gap by providing a framework to generate realistic load consisting of multi-turn conversations with configurable end-user parameters such as number of users in the workload loop, think-time distributions etc. Moreover, LLMRunner provides declarative matrix-sweep experiment plans, integrated GPU and KV-cache monitoring, and an automated artifact collection and plotting subsystem. We use LLMRunner to comparatively analyze the performance of LLM serving when NVIDIA CUDA Unified Virtual Memory (UVM) is used as a drop-in replacement memory allocator for the Key-Value cache (KV-Cache). Our test demonstrates that UVM incurs no measurable throughput or latency penalty compared to the baseline memory allocation technique for non-oversubscription scenarios.

cs.PF

新论文 · v1arXiv:2611.15280

Autonomous Trail Racing with a Bicycle Robot

Fabian Jenelten, Deepana Ishtaweera, Pablo Garcia, Liana Margolese, Igor Bogoslavskyi, Luca Buonocore, Jan Migliazza, Aleksandrs Ecins 等 13 位作者

原文摘要

Navigating off-road trails on a mountain bike requires diverse locomotion skills, including aerial maneuvers, cornering, and slip recovery, paired with sophisticated perception for scene understanding. Translating these capabilities to autonomous, two-wheeled robots remains a formidable challenge. We show that autonomous, high-speed trail navigation can be achieved through a single idea: encoding the trail as a semantic layer of the terrain. In simulation, this layer embeds the navigation task into the training environment; during deployment, it is reconstructed onboard from camera and LiDAR data. First, we introduce a unified learning framework that treats local navigation and locomotion as a joint optimization problem, bypassing the twist or pose command interface of classical control hierarchies. With fixed reward weights, the same framework produces both trail racing and agile parkour, shaped only through terrain semantics and episode length. Second, our perception pipeline combines a frozen vision foundation model with a lightweight dual-head decoder that predicts per-pixel terrain semantics and confidence scores, which are fused with LiDAR range measurements into an egocentric semantic elevation map at 30 Hz onboard an NVIDIA AGX Orin. We validate our framework on the Ultra Mobility Vehicle (UMV), a bicycle robot that autonomously raced across all sections of a kilometer-long, unstructured mountain bike trail at speeds exceeding 5 m/s and averaging 3.26 m/s. In simulation, our policy fails up to an order of magnitude less often than hierarchical baselines while driving faster.

cs.RO

版本更新 · v2arXiv:2604.17861

GPUOS: A GPU Operating System Primitive for Transparent Operation Fusion

Yiwei Yang, Xiangyu Gao, Yuan Zhou, Yuhang Gan, Yusheng Zheng, Andi Quinn

原文摘要

Modern deep learning workloads often consist of many small tensor operations, especially in inference, attention, and micro-batched training. In these settings, kernel launch overhead can become a major bottleneck, sometimes exceeding the actual computation time. We present GPUOS, a GPU runtime JIT system that reduces launch overhead using a persistent kernel architecture with runtime operator injection. GPUOS runs a single long-lived GPU kernel that continuously processes tasks from a host-managed work queue, eliminating repeated kernel launches. To support diverse operations, GPUOS uses NVIDIA NVRTC to just-in-time compile operators at runtime and inject them into the running kernel through device function pointer tables. This design enables operator updates without restarting the kernel or recompiling the system. GPUOS introduces four key ideas: (1) a persistent worker kernel with atomic task queues, (2) a runtime operator injection mechanism based on NVRTC and relocatable device code, (3) a dual-slot aliasing scheme for safe concurrent operator updates, and (4) transparent PyTorch integration through TorchDispatch that batches micro-operations into unified submissions. The system supports arbitrary tensor shapes, strides, data types, and broadcasting through a generic tensor abstraction. Experiments show that GPUOS achieves up to 15.3x speedup over standard PyTorch on workloads dominated by small operations, including micro-batched inference and attention patterns. GPUOS improves utilization while remaining compatible with the PyTorch ecosystem.

cs.DC · cs.OS

新论文 · v1arXiv:2611.14924

Edge-Native Multimodal RAG: Hardware-Aware Quantization and Architectural Optimization for Consumer GPUs

Zahraa Selim, Menna Hamed, Wesam Ahmed, Sohyla Said, Sara Basheer, Rami Zewail

原文摘要

Deployment of retrieval-augmented generation (RAG) for Edge-AI is constrained by memory and bandwidth limitations. A standard 7B parameter model consumes 14 GB Video Random Access Memory (VRAM) in half-precision (FP16), precluding co-location of vector indices and Key-Value (KV) caches. This paper presents an engineering integration and empirical evaluation of a multimodal RAG framework optimized for consumer-grade Turing hardware (e.g., NVIDIA Tesla T4). Our system-level analysis identifies a critical trade-off: while generative pre-trained transformer quantization (GPTQ) offers superior throughput (51.6 tok/s), it exhibits significant context instability at 2048+ tokens on Turing architectures. Conversely, NormalFloat 4-bit (NF4) maintains robust long-context performance with viable latency (81 ms/tok). By coupling NF4 with a decoupled Florence-2 vision integration, we achieve visual reasoning with only 430 MB VRAM overhead. Furthermore, empirical observations of prompt-induced search space pruning show that structured prompting reduces latency by 47% without altering computational intensity (1.88 GFLOPs/token). The integrated architecture maintains an 8.37 GB footprint, establishing a viable RAG pipeline for 12-16 GB GPUs. Finally, structured few-shot prompting effectively compensates for quantization-induced reasoning loss, restoring mathematical accuracy from 35.0% to 92.0% across 200 trials.

cs.AI

版本更新 · v2arXiv:2607.14203

Instant NuRec: Feed-Forward 3D Gaussian Reconstruction for Driving Scene Simulation

NVIDIA, :, Jiahui Huang, Jiawei Ren, Michal Tyszkiewicz, Bjoern Haefner, Michael Shelley, Xin Kang 等 17 位作者

原文摘要

3D simulation platforms are critical for autonomous driving because they enable end-to-end policy evaluation, thereby reducing development costs and improving safety. In recent years, neural simulation has become predominant, with methods such as NuRec playing a central role; however, these methods remain relatively slow and typically require per-scene tuning. In this work, we present Instant NuRec, a feed-forward neural reconstruction model that turns a short multi-view driving log into a fully simulatable 3D Gaussian Splatting (3DGS) world in a single forward pass. The model accepts multi-view input from a calibrated camera rig and emits a layered output consisting of static and dynamic 3DGS layers, a sky cubemap, and per-camera ISP corrections, while providing native support for non-pinhole camera models via 3DGUT. It reconstructs a 10-20-second multi-camera scene in roughly 1.5 seconds and achieves a PSNR on the Waymo Open Dataset that is 2.01 dB above the strongest evaluated baseline. Instant NuRec is deeply integrated into NuRec and is compatible with AlpaSim for closed-loop simulation.

cs.GR · cs.AI · cs.CV

新论文 · v1arXiv:2611.14815

BirdDreams: Joint Perspective and BEV Video Generation for a Plausible, Reactive Driving World Model

Junyoung Seo, Jaewoo Seo, Seungryong Kim, Laura Leal-Taixé, Xiuming Zhang, Seung Wook Kim

原文摘要

Driving world models aim to simulate how a scene unfolds under different ego actions. This requires plausible motion of surrounding vehicles as well as precise ego-trajectory control. Learning vehicle motion from perspective video alone is challenging because motion is entangled with appearance and vehicles can be occluded. Prior work therefore conditions video generation on layouts that specify vehicle motion, providing control over surrounding vehicles but requiring their future trajectories as inputs. We introduce BirdDreams, an autoregressive driving world model that jointly generates perspective video and a schematic bird's-eye-view (BEV) video, making scene layout part of the predicted future. The BEV, rendered from scene annotations for training, depicts roads and vehicles as simple shapes at a fixed metric scale, providing a top-down view of their positions and motion with reduced occlusion between vehicles. Perspective video, by contrast, retains scene details that the BEV omits, such as traffic signals and object appearance. Joint attention between two video diffusion transformers, one per stream, lets each stream inform the other during generation. The predicted BEV also provides a geometric interface for ego control: we translate and rotate its latest frame along an input ego trajectory and use the resulting frames as a soft condition that shows how the scene would appear to move if only the ego vehicle moved. At inference, BirdDreams generates both videos from a single camera image without requiring an HD map or vehicle-state inputs. On PhysicalAI-AV and nuScenes, BirdDreams follows input ego trajectories more faithfully than prior driving world models and generates more plausible ego motion without ego-action conditioning. Human evaluation shows that BirdDreams outperforms a perspective-only baseline in generating realistic surrounding-vehicle behavior.

cs.CV

新论文 · v1arXiv:2611.14741

Reconstructing Fine-grained Memory Reuse Structure in GPU Kernels

Yanbo Zhao, Yueming Hao, Zecheng Li, Mao Lin, Shuyin Jiao, Keren Zhou, Xu Liu, Jiajia Li

原文摘要

Optimizing GPU data movement requires knowing which instructions access the same data and across which execution scopes reuse occurs. Existing profilers can attribute memory behavior to individual instructions but do not expose the reuse relations among them, while aliasing, dynamic indexing, and runtime parameters prevent static analysis from establishing these relations in general. Cycle-level simulation can reconstruct such relations, but at costs prohibitive for production workloads. We introduce two abstractions for representing this structure: Topological Distance, which labels each dynamic memory access with the smallest execution scope (thread, warp, block, or grid) containing its most recent reuse, and PCDepGraph, which organizes these labels into a PC-level graph of the kernel's memory reuse structure. Defined over logical execution scopes rather than transient hardware state, the representation remains stable across runs and GPU targets while directly indicating the sharing scope that candidate optimization mechanisms must cover. Our tool, cuVein, reconstructs PCDepGraph directly from optimized binaries, without sampling, at generally under 10x overhead. Across HPC and AI workloads on NVIDIA RTX 4090 and H100, optimizations guided by PCDepGraph eliminate up to 90.18% of memory instructions and yield up to 2.84x kernel speedup; reuse-scope distributions remain consistent across platforms, supporting that the representation captures program structure rather than transient hardware behavior.

cs.DC · cs.PF · cs.PL

新论文 · v1arXiv:2611.14158

Neural network scaling and generalization for 3D strain fields in heterogeneous materials

Lucas Fourel, Paul Devianne, Marine Debouvry, Tristan Destruhaut, Russell Gunn, Ashwini Gupta, Lori Graham-Brady, Jean-François Molinari

原文摘要

Computational homogenization and coupled multiscale simulations of heterogeneous materials require repeated solution of mechanical fields on Representative Volume Elements, which becomes prohibitively expensive for high-resolution three-dimensional microstructures. Machine-learning surrogates can accelerate these computations, yet most existing approaches predict fields for fixed macroscopic loadings and provide limited guidance on how to scale network architectures for high-resolution 3D tensor-field prediction. This work presents a U-Net surrogate trained to predict the full 36-component strain localization tensor field in voxelized concrete mesostructures at resolutions up to $256^3$. Training localization tensors are generated with an FFT-based linear elastic solver, and an extensive hyperparameter optimization is performed to derive scaling rules and an accuracy-speed trade-off analysis. The results show that only network width and depth need to be scaled with problem complexity, while several conventional U-Net design choices, such as doubling channels after downsampling, are suboptimal for this problem. The final neural network achieves a Mean Absolute Error of 1.118% on both the validation and the held-out test sets, and an inference speedup of approximately 8.7x over the FFT-based solver on an Nvidia H100 GPU. Generalization experiments reveal transfer across aggregate volume fractions and highlight the importance of training on diverse morphologies. To safeguard reliability, a Maximum Mean Discrepancy metric computed on internal feature maps is proposed for detecting out-of-distribution microstructures without costly reference computations.

physics.comp-ph · cs.CE

新论文 · v1arXiv:2611.14083

Sable: A Small-Field Sumcheck Argument for Bit-Exact Inference of Large Language Models

Tomer Solberg

原文摘要

We present Sable, an argument system and single-GPU prover for the statement that a committed, integer-quantized transformer produced a given token sequence from a given prompt, bit for bit. Sable fixes an integer function F_O with no floating-point step, in which every rounding, normalization, softmax and sampling decision is determined by its inputs, and proves it with a constraint system whose relations provably never wrap modulo the 31-bit KoalaBear prime. The proof batches zerochecks and LogUp-GKR lookups into one sumcheck per domain, and commits to every witness tensor and weight in one multilinear commitment built from Reed-Solomon codes, BLAKE3 Merkle trees and a mixed-length BaseFold opening. The prover streams the model layer by layer on one GPU. On one NVIDIA RTX PRO 6000 Blackwell GPU, Sable proves all 64 layers of Qwen2.5-32B in 92.5 s for 4,096 tokens and 398 s for 16,384 tokens, and covers Qwen3-32B and the mixture-of-experts Qwen3-30B-A3B (4,096 tokens in 121.8 and 84.7 s). A proof costs 41 to 50 times the serving time, or 6.4 cents per 4,096-token request at public GPU prices, and 40 to 64 times less time per parameter and token than zkLLM. On WikiText-2, F_O's greedy token matches the BF16 model's at 86.3% of positions; on Qwen3, F_O is within 1.6 points of BF16 on MMLU and 1.7 points on GSM8K. A verifier with a compiled core checks a proof on one CPU core in 19 to 56 s. In the random-oracle model, the interactive proof has round-by-round soundness error 2^-103. The proof is not zero-knowledge, so the system targets open-weight models. Code: https://github.com/moonmath-ai/Sable

cs.CR · cs.AI · cs.LG

版本更新 · v3arXiv:2603.16382

Rotated Robustness: A Training-Free Defense against Bit-Flip Attacks on Large Language Models

Deng Liu, Song Chen

原文摘要

Bit-flip corruption of quantized weights poses a serious reliability threat to Large Language Models (LLMs), as even a small number of weight faults can trigger catastrophic model degradation. We show that such failures can be strongly amplified when corrupted weights are coupled to high-sensitivity activation channels. Based on this observation, we propose Rotated Robustness (RoR), a training-free sensitivity-aware hierarchical rotation that applies matched orthogonal transformations to activations and weights, redistributing sensitive channel contributions across feature dimensions while preserving the underlying linear mapping in exact arithmetic. Across eight evaluated LLMs, RoR matches or exceeds all compared methods in the number of cumulative PBS flips sustained before PPL exceeds 100 within the 100-flip evaluation horizon. Under BER-based random bit-flip injection, no RoR failures are observed in the evaluated 1,000-trial experiments for the tested models. Under defense-aware SPoF reconstruction, RoR increases the cost of reproducing specified one-bit catastrophic failures to thousands of deployed INT8 bit changes. RoR also preserves downstream-task utility under PBS perturbations. On NVIDIA H200 at batch size 8, RoR incurs 0.097\%--0.584\% persistent storage overhead, 5.2\%--14.1\% Prefill overhead, and 0.5\%--9.8\% Decode overhead across four representative models. These results show that sensitivity-aware hierarchical rotation provides a practical approach to improving the bit-flip robustness of quantized LLM Linear weights.

cs.CR

新论文 · v1arXiv:2611.13879

SFU-Sim: Bit-Exact Reverse Engineering of NVIDIA MUFU and Validation of PTX Paths

Haozhi Xie, Tao Su

原文摘要

Reproducing the numerical behavior of a GPU requires predicting the exact output of special function instructions. Vendor documents give only the function semantics and error bounds. These do not fix the output bit pattern. We present SFU-Sim, which builds bit-exact models of NVIDIA multi-function unit (MUFU) instructions by black-box reverse engineering on real GPUs. We first model a single bare MUFU instruction. We then add the surrounding operations found in the disassembly to model full PTX paths. The model covers 12 common MUFU forms and two BF16 forms on the tested Blackwell GPU. Using SIN/COS as a complete case study, we show how we analyze the PTX-to-SASS compilation path and recover the segment structure, the equivalent coefficient tables, and the finite-width arithmetic. On NVIDIA GPUs of three architectures (Ampere, Ada Lovelace, and Blackwell), we exhaustively compare every tested PTX path over its specified input domain. FP16 and FP32 paths cover all $2^{16}$ and $2^{32}$ input encodings, respectively. The model output matches the hardware bit for bit, with zero mismatches. SFU-Sim computes each output directly from its input. It needs neither a lookup table of all inputs nor access to a physical GPU. It can serve as a bit-exact special function component for instruction simulators. It can also serve as an arithmetic reference for hardware design and functional verification that target the same numerical behavior. SFU-Sim and its validation programs are open source at https://github.com/hzX40000000/SFU-Sim

cs.AR

新论文 · v1arXiv:2611.13860

Uncovering the Memory Hierarchy of Modern NVIDIA GPUs for Accurate Simulation

Ziwen Wang, Jiapeng Zhou, Changxi Liu, Dongsheng Wang, Yun Chen

原文摘要

Emerging artificial intelligence workloads, particularly Large Language Models (LLMs), place unprecedented demands on GPU memory systems. Consequently, architectural innovations increasingly target the memory system to alleviate these bottlenecks. However, existing works still rely on outdated memory hierarchy models that fail to capture the complexities of modern hardware. This fidelity gap leads to inaccurate characterization and could misguide optimization strategies. To bridge this gap, we present GemAtlas, the first systematic reverse engineering of the memory hierarchy in modern NVIDIA GPUs. By developing fine-grained SASS-level benchmarking, we expose critical, previously undocumented micro-architectural details across different cache levels, including cache set associativity, slice-routing and set-indexing hash functions, replacement policies, coherence behavior, and interconnect topologies. We integrate these discovered micro-architecture details into the state-of-the-art GPU simulation framework. The final results demonstrate that GemAtlas reduces the mean absolute percentage error (MAPE) from 31.69% to 22.24% on average across four GPUs. Specifically, GemAtlas reduces the MAPE from 34.04% to 23.98% for real-world LLM operators.

cs.AR