{"total":115,"rows":[{"id":"2205.14135","title":"FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness","created":"2022-05-27","cited_by_count":3673,"area":null},{"id":"2307.08691","title":"FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning","created":"2023-07-17","cited_by_count":2598,"area":null},{"id":"2310.01889","title":"Ring Attention with Blockwise Transformers for Near-Infinite Context","created":"2023-10-03","cited_by_count":457,"area":null},{"id":"2406.16838","title":"From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Models","created":"2024-06-24","cited_by_count":137,"area":null},{"id":"2502.18137","title":"SpargeAttention: Accurate and Training-free Sparse Attention Accelerating Any Model Inference","created":"2025-02-25","cited_by_count":120,"area":null},{"id":"2411.10958","title":"SageAttention2: Efficient Attention with Thorough Outlier Smoothing and Per-thread INT4 Quantization","created":"2024-11-17","cited_by_count":114,"area":null},{"id":"2403.02308","title":"Vision-RWKV: Efficient and Scalable Visual Perception with RWKV-Like Architectures","created":"2024-03-04","cited_by_count":98,"area":null},{"id":"2302.14017","title":"Full Stack Optimization of Transformer Inference: a Survey","created":"2023-02-27","cited_by_count":80,"area":null},{"id":"2311.12351","title":"Advancing Transformer Architecture in Long-Context Large Language Models: A Comprehensive Survey","created":"2023-11-21","cited_by_count":72,"area":null},{"id":"2505.11594","title":"SageAttention3: Microscaling FP4 Attention for Inference and An Exploration of 8-Bit Training","created":"2025-05-16","cited_by_count":66,"area":null},{"id":"2603.15031","title":"Attention Residuals","created":"2026-03-16","cited_by_count":65,"area":null},{"id":"2411.09009","title":"Cut Your Losses in Large-Vocabulary Language Models","created":"2024-11-13","cited_by_count":40,"area":null},{"id":"2505.16381","title":"PaTH Attention: Position Encoding via Accumulating Householder Transformations","created":"2025-05-22","cited_by_count":34,"area":null},{"id":"2311.05908","title":"FlashFFTConv: Efficient Convolutions for Long Sequences with Tensor Cores","created":"2023-11-10","cited_by_count":33,"area":null},{"id":"2412.03594","title":"BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching","created":"2024-11-29","cited_by_count":32,"area":null},{"id":"2310.03294","title":"DISTFLASHATTN: Distributed Memory-efficient Attention for Long-context LLMs Training","created":"2023-10-05","cited_by_count":29,"area":null},{"id":"2411.01783","title":"Context Parallelism for Scalable Million-Token Inference","created":"2024-11-04","cited_by_count":26,"area":null},{"id":"2302.04542","title":"Efficient Attention via Control Variates","created":"2023-02-09","cited_by_count":23,"area":null},{"id":"2505.21136","title":"SageAttention2++: A More Efficient Implementation of SageAttention2","created":"2025-05-27","cited_by_count":22,"area":null},{"id":"2410.01359","title":"FlashMask: Efficient and Rich Mask Extension of FlashAttention","created":"2024-10-02","cited_by_count":20,"area":null},{"id":"2105.12842","title":"A Full-Stack Search Technique for Domain Optimized Deep Learning Accelerators","created":"2021-05-26","cited_by_count":19,"area":null},{"id":"2504.16922","title":"Generalized Neighborhood Attention: Multi-dimensional Sparse Attention at the Speed of Light","created":"2025-04-23","cited_by_count":18,"area":null},{"id":"2405.10480","title":"Lean Attention: Hardware-Aware Scalable Attention Mechanism for the Decode-Phase of Transformers","created":"2024-05-17","cited_by_count":16,"area":null},{"id":"2312.03134","title":"A Hardware Evaluation Framework for Large Language Model Inference","created":"2023-12-05","cited_by_count":14,"area":null},{"id":"2403.04690","title":"Faster Neighborhood Attention: Reducing the O(n^2) Cost of Self Attention at the Threadblock Level","created":"2024-03-07","cited_by_count":14,"area":null},{"id":"2402.14808","title":"RelayAttention for Efficient Large Language Model Serving with Long System Prompts","created":"2024-02-22","cited_by_count":13,"area":null},{"id":"2503.14376","title":"Tiled Flash Linear Attention: More Efficient Linear RNN and xLSTM Kernels","created":"2025-03-18","cited_by_count":13,"area":null},{"id":"2305.19370","title":"Blockwise Parallel Transformer for Large Context Models","created":"2023-05-30","cited_by_count":12,"area":null},{"id":"2502.12082","title":"AdaSplash: Adaptive Sparse Flash Attention","created":"2025-02-17","cited_by_count":12,"area":null},{"id":"2505.12044","title":"FlashBias: Fast Computation of Attention with Bias","created":"2025-05-17","cited_by_count":11,"area":null},{"id":"2602.03067","title":"FlashSinkhorn: IO-Aware Entropic Optimal Transport on GPU","created":"2026-02-03","cited_by_count":11,"area":null},{"id":"2403.09347","title":"BurstAttention: An Efficient Distributed Attention Framework for Extremely Long Sequences","created":"2024-03-14","cited_by_count":10,"area":null},{"id":"2409.16997","title":"INT-FlashAttention: Enabling Flash Attention for INT8 Quantization","created":"2024-09-25","cited_by_count":9,"area":null},{"id":"2404.02945","title":"Optimizing the Deployment of Tiny Transformers on Low-Power MCUs","created":"2024-04-03","cited_by_count":8,"area":null},{"id":"2508.08435","title":"Fast weight programming and linear transformers: from machine learning to neurobiology","created":"2025-08-11","cited_by_count":8,"area":null},{"id":"2508.18224","title":"FSA: An Alternative Efficient Implementation of Native Sparse Attention Kernel","created":"2025-08-25","cited_by_count":7,"area":null},{"id":"2407.08330","title":"HDT: Hierarchical Document Transformer","created":"2024-07-11","cited_by_count":6,"area":null},{"id":"2408.16978","title":"Training Ultra Long Context Language Model with Fully Pipelined Distributed Transformer","created":"2024-08-30","cited_by_count":6,"area":null},{"id":"2410.08481","title":"Generation with Dynamic Vocabulary","created":"2024-10-11","cited_by_count":6,"area":null},{"id":"2504.14519","title":"SlimPipe: Memory-Thrifty and Efficient Pipeline Parallelism for Long-Context LLM Training","created":"2025-04-20","cited_by_count":6,"area":null},{"id":"2312.11918","title":"A Case Study in CUDA Kernel Fusion: Implementing FlashAttention-2 on NVIDIA Hopper Architecture using the CUTLASS Library","created":"2023-12-19","cited_by_count":5,"area":null},{"id":"2505.14314","title":"Low-Cost FlashAttention with Fused Exponential and Multiplication Hardware Operators","created":"2025-05-20","cited_by_count":5,"area":null},{"id":"2510.17189","title":"SOLE: Hardware-Software Co-design of Softmax and LayerNorm for Efficient Transformer Inference","created":"2025-10-20","cited_by_count":5,"area":null},{"id":"2409.15104","title":"PecSched: Preemptive and Efficient Cluster Scheduling for LLM Inference","created":"2024-09-23","cited_by_count":4,"area":null},{"id":"2411.17847","title":"SoftmAP: Software-Hardware Co-design for Integer-Only Softmax on Associative Processors","created":"2024-11-26","cited_by_count":4,"area":null},{"id":"2501.14256","title":"DKT2: Revisiting Applicable and Comprehensive Knowledge Tracing in Large-Scale Data","created":"2025-01-24","cited_by_count":4,"area":null},{"id":"2505.23520","title":"AnchorAttention: Difference-Aware Sparse Attention with Stripe Granularity","created":"2025-05-29","cited_by_count":4,"area":null},{"id":"2510.01450","title":"Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regression","created":"2025-10-01","cited_by_count":4,"area":null},{"id":"2105.02103","title":"Prototype Memory for Large-scale Face Representation Learning","created":"2021-05-05","cited_by_count":3,"area":null},{"id":"2402.10118","title":"Reusing Softmax Hardware Unit for GELU Computation in Transformers","created":"2024-02-15","cited_by_count":3,"area":null},{"id":"2410.16663","title":"FastAttention: Extend FlashAttention2 to NPUs and Low-resource GPUs","created":"2024-10-22","cited_by_count":3,"area":null},{"id":"2503.00461","title":"Leveraging Compute-in-Memory for Efficient Generative Model Inference in TPUs","created":"2025-03-01","cited_by_count":3,"area":null},{"id":"2505.22758","title":"FlashFormer: Whole-Model Kernels for Efficient Low-Batch Inference","created":"2025-05-28","cited_by_count":3,"area":null},{"id":"2507.04239","title":"Scaling Context Requires Rethinking Attention","created":"2025-07-06","cited_by_count":3,"area":null},{"id":"2507.04416","title":"RAT: Bridging RNN Efficiency and Attention Accuracy via Chunk-based Sequence Modeling","created":"2025-07-06","cited_by_count":3,"area":null},{"id":"2511.13274","title":"KForge: Program Synthesis for Diverse AI Hardware Accelerators","created":"2025-11-17","cited_by_count":3,"area":null},{"id":"2603.15854","title":"FlashSampling: Fast and Memory-Efficient Exact Sampling","created":"2026-03-16","cited_by_count":3,"area":null},{"id":"2604.15180","title":"AdaSplash-2: Faster Differentiable Sparse Attention","created":"2026-04-16","cited_by_count":3,"area":null},{"id":"2505.14201","title":"FLASH-D: FlashAttention with Hidden Softmax Division","created":"2025-05-20","cited_by_count":2,"area":null},{"id":"2506.21137","title":"Norm$\\times$Direction: Restoring the Missing Query Norm in Vision Linear Attention","created":"2025-06-26","cited_by_count":2,"area":null},{"id":"2507.16676","title":"Custom Algorithm-based Fault Tolerance for Attention Layers in Transformers","created":"2025-07-22","cited_by_count":2,"area":null},{"id":"2511.02168","title":"Eliminating Multi-GPU Performance Taxes: A Systems Approach to Efficient Distributed LLMs","created":"2025-11-04","cited_by_count":2,"area":null},{"id":"2602.06063","title":"Mapping Gemma3 onto an Edge Dataflow Architecture","created":"2026-01-27","cited_by_count":2,"area":null},{"id":"2604.03425","title":"AEGIS: Scaling Long-Sequence Homomorphic Encrypted Transformer Inference via Hybrid Parallelism on Multi-GPU Systems","created":"2026-04-03","cited_by_count":2,"area":null},{"id":"2604.12798","title":"VFA: Relieving Vector Operations in Flash Attention with Global Maximum Pre-computation","created":"2026-04-14","cited_by_count":2,"area":null},{"id":"2606.22984","title":"Scalable Physics-Inspired Transformers for Spin Glasses","created":"2026-06-22","cited_by_count":2,"area":null},{"id":"2607.26694","title":"Visko Orbis 1.0: A Live Model for Real-Time Interactive Long Video Generation","created":"2026-07-29","cited_by_count":2,"area":null},{"id":"2407.02109","title":"HRSAM: Efficient Interactive Segmentation in High-Resolution Images","created":"2024-07-02","cited_by_count":1,"area":null},{"id":"2409.15373","title":"Enhancing Performance and Scalability of Large-Scale Recommendation Systems with Jagged Flash Attention","created":"2024-09-19","cited_by_count":1,"area":null},{"id":"2409.19134","title":"Confidential Prompting: Privacy-preserving LLM Inference on Cloud","created":"2024-09-27","cited_by_count":1,"area":null},{"id":"2412.16481","title":"Flash3D: Super-scaling Point Transformers through Joint Hardware-Geometry Locality","created":"2024-12-21","cited_by_count":1,"area":null},{"id":"2502.01659","title":"Longer Attention Span: Increasing Transformer Context Length with Sparse Graph Processing Techniques","created":"2025-01-31","cited_by_count":1,"area":null},{"id":"2503.14985","title":"ML-Triton, A Multi-Level Compilation and Language Extension to Triton GPU Programming","created":"2025-03-19","cited_by_count":1,"area":null},{"id":"2506.20686","title":"MegaFold: Efficient Training of Next-Generation 3D Attention Protein Models on Cross-Platform GPUs","created":"2025-06-24","cited_by_count":1,"area":null},{"id":"2511.17599","title":"From Projection to Prediction: Beyond Logits for Scalable Language Models","created":"2025-11-18","cited_by_count":1,"area":null},{"id":"2602.22575","title":"S2O: Early Stopping for Sparse Attention via Online Permutation","created":"2026-02-26","cited_by_count":1,"area":null},{"id":"2605.29157","title":"Parallax: Parameterized Local Linear Attention for Language Modeling","created":"2026-05-27","cited_by_count":1,"area":null},{"id":"2606.24780","title":"BluTrain: A C++/CUDA Framework for AI Systems","created":"2026-06-23","cited_by_count":1,"area":null},{"id":"2607.12789","title":"AVQ-Attention: Adaptive Vector-Quantized Attention","created":"2026-07-14","cited_by_count":1,"area":null},{"id":"2607.25291","title":"CoSA: Accelerating Long-Context Inference via Proxy-Kernel Co-Designed Sparse Attention","created":"2026-07-28","cited_by_count":1,"area":null},{"id":"2405.04206","title":"NOVA: NoC-based Vector Unit for Mapping Attention Layers on a CNN Accelerator","created":"2024-05-07","cited_by_count":0,"area":null},{"id":"2503.15758","title":"ATTENTION2D: Communication Efficient Distributed Self-Attention Mechanism","created":"2025-03-20","cited_by_count":0,"area":null},{"id":"2510.06834","title":"Vectorized FlashAttention with Low-cost Exponential Computation in RISC-V Vector Processors","created":"2025-10-08","cited_by_count":0,"area":null},{"id":"2511.02043","title":"Flashlight: PyTorch Compiler Extensions to Accelerate Attention Variants","created":"2025-11-03","cited_by_count":0,"area":null},{"id":"2603.06731","title":"PolyBlocks: A Compiler Infrastructure for AI Chips and Programming Frameworks","created":"2026-03-06","cited_by_count":0,"area":null},{"id":"2604.20819","title":"Stream-CQSA: Exact Out-of-Memory Recovery for Attention","created":"2026-04-22","cited_by_count":0,"area":null},{"id":"2604.23466","title":"Evaluating CUDA Tile for AI Workloads on Hopper and Blackwell GPUs","created":"2026-04-25","cited_by_count":0,"area":null},{"id":"2604.23798","title":"ELSA: Exact Linear-Scan Attention for Fast and Memory-Light Vision Transformers","created":"2026-04-26","cited_by_count":0,"area":null},{"id":"2604.24715","title":"Long-Context Aware Upcycling: A New Frontier for Hybrid LLM Scaling","created":"2026-04-27","cited_by_count":0,"area":null},{"id":"2604.24820","title":"Salca: A Sparsity-Aware Hardware Accelerator for Efficient Long-Context Attention Decoding","created":"2026-04-27","cited_by_count":0,"area":null},{"id":"2604.25306","title":"QFlash: Bridging Quantization and Memory Efficiency in Vision Transformer Attention","created":"2026-04-28","cited_by_count":0,"area":null},{"id":"2606.02963","title":"KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators","created":"2026-06-01","cited_by_count":0,"area":null},{"id":"2606.30460","title":"HSAP: A Hierarchical Sequence-aware Parallelism for Hybrid-Context Generative Models","created":"2026-06-29","cited_by_count":0,"area":null},{"id":"2607.05098","title":"Directly Optimizing Mean Demographic Parity for Nonlinear Regression","created":"2026-07-06","cited_by_count":0,"area":null},{"id":"2607.05439","title":"Design-CP: Context Parallelism for Design of Protein Nanoparticles","created":"2026-07-03","cited_by_count":0,"area":null},{"id":"2607.09385","title":"STEEL: Sparsity-Aware Fused Attention for Energy-Efficient Long-Sequence Inference on AMD's XDNA NPU","created":"2026-07-10","cited_by_count":0,"area":null},{"id":"2607.20457","title":"Dropping the Anchor: Statistical Context Summarization for Distributed Systems via Pulsar Attention","created":"2026-05-14","cited_by_count":0,"area":null},{"id":"2607.26988","title":"A Compositional Theory of Causally Masked Transformers","created":"2026-07-29","cited_by_count":0,"area":null},{"id":"2608.02091","title":"One QK Channel, Many Sources: Tracing Low-Precision Attention Collapse","created":"2026-08-03","cited_by_count":0,"area":null},{"id":"2608.07616","title":"HSMLA: Hierarchical Softmax Multi-scale Linear Attention for Efficient Vision Transformers","created":"2026-08-07","cited_by_count":0,"area":null}],"next_cursor":"100"}