Copertina

Sequence-Aware Split Heuristic to Mitigate SM Underutilization in FlashAttention-3 Low-Head-Count Decoding

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Font, Martí Llopart, Hernando, Javier, España-Bonet, Cristina
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Hardware Architecture Distributed, Parallel, and Cluster Computing
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Accelerating Triangle Counting with Real Processing-in-Memory Systems
di: Asquini, Lorenzo, et al.
Pubblicazione: (2025)

Mitigating Shared Storage Congestion Using Control Theory
di: Collignon, Thomas, et al.
Pubblicazione: (2025)

DUET: Disaggregated Hybrid Mamba-Transformer LLMs with Prefill and Decode-Specific Packages
di: Kanani, Alish, et al.
Pubblicazione: (2026)

SAGe: A Lightweight Algorithm-Architecture Co-Design for Mitigating the Data Preparation Bottleneck in Large-Scale Genome Sequence Analysis
di: Ghiasi, Nika Mansouri, et al.
Pubblicazione: (2025)

On-Package Memory with Universal Chiplet Interconnect Express (UCIe): A Low Power, High Bandwidth, Low Latency and Low Cost Approach
di: Sharma, Debendra Das, et al.
Pubblicazione: (2025)

HieraSparse: Hierarchical Semi-Structured Sparse KV Attention
di: Wang, Haoxuan, et al.
Pubblicazione: (2026)

Part-time Power Measurements: nvidia-smi's Lack of Attention
di: Yang, Zeyu, et al.
Pubblicazione: (2023)

SnapStream: Efficient Long Sequence Decoding on Dataflow Accelerators
di: Li, Jonathan, et al.
Pubblicazione: (2025)

Knowledge-Guided Attention-Inspired Learning for Task Offloading in Vehicle Edge Computing
di: Ma, Ke, et al.
Pubblicazione: (2025)

Infinite-LLM: Efficient LLM Service for Long Context with DistAttention and Distributed KVCache
di: Lin, Bin, et al.
Pubblicazione: (2024)

Optimizing Task Scheduling in Fog Computing with Deadline Awareness
di: Sirjani, Mohammad Sadegh, et al.
Pubblicazione: (2025)

Workload-Aware Hardware Accelerator Mining for Distributed Deep Learning Training
di: Adnan, Muhammad, et al.
Pubblicazione: (2024)

Simopt-Power: Leveraging Simulation Metadata for Low-Power Design Synthesis
di: Wadhwa, Eashan, et al.
Pubblicazione: (2025)

Enabling Time-Aware Priority Traffic Management over Distributed FPGA Nodes
di: Scionti, Alberto, et al.
Pubblicazione: (2025)

Topology-Aware Virtualization over Inter-Core Connected Neural Processing Units
di: Feng, Dahu, et al.
Pubblicazione: (2025)

Towards Compute-Aware In-Switch Computing for LLMs Tensor-Parallelism on Multi-GPU Systems
di: Zhang, Chen, et al.
Pubblicazione: (2026)

MVDRAM: Enabling GeMV Execution in Unmodified DRAM for Low-Bit LLM Acceleration
di: Kubo, Tatsuya, et al.
Pubblicazione: (2025)

Cache Your Prompt When It's Green: Carbon-Aware Caching for Large Language Model Serving
di: Tian, Yuyang, et al.
Pubblicazione: (2025)

TeraPool: A Physical Design Aware, 1024 RISC-V Cores Shared-L1-Memory Scaled-up Cluster Design with High Bandwidth Main Memory Link
di: Zhang, Yichao, et al.
Pubblicazione: (2026)

FlashMoE: Fast Distributed MoE in a Single Kernel
di: Aimuyo, Osayamen Jonathan, et al.
Pubblicazione: (2025)

PIMDAL: Mitigating the Memory Bottleneck in Data Analytics using a Real Processing-in-Memory System
di: Frouzakis, Manos, et al.
Pubblicazione: (2025)

CIPHERMATCH: Accelerating Homomorphic Encryption-Based String Matching via Memory-Efficient Data Packing and In-Flash Processing
di: Kabra, Mayank, et al.
Pubblicazione: (2025)

MegIS: High-Performance, Energy-Efficient, and Low-Cost Metagenomic Analysis with In-Storage Processing
di: Ghiasi, Nika Mansouri, et al.
Pubblicazione: (2024)

TT-Edge: A Hardware-Software Co-Design for Energy-Efficient Tensor-Train Decomposition on Edge AI
di: Kwak, Hyunseok, et al.
Pubblicazione: (2025)

The DEEP-ER project: I/O and resiliency extensions for the Cluster-Booster architecture
di: Kreuzer, Anke, et al.
Pubblicazione: (2019)

An Evaluation and Comparison of GPU Hardware and Solver Libraries for Accelerating the OPM Flow Reservoir Simulator
di: Qiu, Tong Dong, et al.
Pubblicazione: (2023)

SLIM: A Heterogeneous Accelerator for Edge Inference of Sparse Large Language Model via Adaptive Thresholding
di: Xu, Weihong, et al.
Pubblicazione: (2025)

COMET: A Framework for Modeling Compound Operation Dataflows with Explicit Collectives
di: Negi, Shubham, et al.
Pubblicazione: (2025)

PAM: Processing Across Memory Hierarchy for Efficient KV-centric LLM Serving System
di: Liu, Lian, et al.
Pubblicazione: (2026)

RAPID-Graph: Recursive All-Pairs Shortest Paths Using Processing-in-Memory for Dynamic Programming on Graphs
di: Chen, Yanru, et al.
Pubblicazione: (2025)

Chopper: A Multi-Level GPU Characterization Tool & Derived Insights Into LLM Training Inefficiency
di: Kurzynski, Marco, et al.
Pubblicazione: (2025)

Efficient deadlock avoidance for 2D mesh NoCs that use OQ or VOQ routers
di: Papaphilippou, Philippos, et al.
Pubblicazione: (2023)

DCRA: A Distributed Chiplet-based Reconfigurable Architecture for Irregular Applications
di: Orenes-Vera, Marcelo, et al.
Pubblicazione: (2023)

LFOC: A Lightweight Fairness-Oriented Cache Clustering Policy for Commodity Multicores
di: García-García, Adrián, et al.
Pubblicazione: (2024)

FlexVector: A SpMM Vector Processor with Flexible VRF for GCNs on Varying-Sparsity Graphs
di: Li, Bohan, et al.
Pubblicazione: (2026)

iHAC: A Hybrid Cluster Architecture for Enhanced Performance and Resilience
di: Muntaka, Siddique Abubakr, et al.
Pubblicazione: (2026)

NetSmith: An Optimization Framework for Machine-Discovered Network Topologies
di: Green, Conor, et al.
Pubblicazione: (2024)

SpArch: Efficient Architecture for Sparse Matrix Multiplication
di: Zhang, Zhekai, et al.
Pubblicazione: (2020)

Navigating the Landscape of Distributed File Systems: Architectures, Implementations, and Considerations
di: Pan, Xueting, et al.
Pubblicazione: (2024)

Efficient MoE Serving in the Memory-Bound Regime: Balance Activated Experts, Not Tokens
di: Yu, Yanpeng, et al.
Pubblicazione: (2025)