惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

WordPress大学
WordPress大学
G
Google Developers Blog
小众软件
小众软件
V
V2EX
月光博客
月光博客
腾讯CDC
aimingoo的专栏
aimingoo的专栏
J
Java Code Geeks
Y
Y Combinator Blog
人人都是产品经理
人人都是产品经理
B
Blog RSS Feed
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
Microsoft Azure Blog
Microsoft Azure Blog
博客园 - 【当耐特】
D
Docker
M
MIT News - Artificial intelligence
Google DeepMind News
Google DeepMind News
N
Netflix TechBlog - Medium
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
I
InfoQ
MongoDB | Blog
MongoDB | Blog
Apple Machine Learning Research
Apple Machine Learning Research
Jina AI
Jina AI

cs.DC updates on arXiv.org

DUAL-BLADE: Dual-Path NVMe-Direct KV-Cache Offloading for Edge LLM Inference Progressive Semantic Communication for Efficient Edge-Cloud Vision-Language Models Efficient, VRAM-Constrained xLM Inference on Clients Folding Tensor and Sequence Parallelism for Memory-Efficient Transformer Training & Inference DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving RaMP: Runtime-Aware Megakernel Polymorphism for Mixture-of-Experts Spark Policy Toolkit: Semantic Contracts and Scalable Execution for Policy Learning in Spark Internet of Everything in the 6G Era: Paradigms, Enablers, Potentials and Future Directions PolyKV: A Shared Asymmetrically-Compressed KV Cache Pool for Multi-Agent LLM Inference A Survey on Split Learning for LLM Fine-Tuning: Models, Systems, and Privacy Optimizations ITAS: A Multi-Agent Architecture for LLM-Based Intelligent Tutoring Latency and Cost of Multi-Agent Intelligent Tutoring at Scale TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training FreeScale: Distributed Training for Sequence Recommendation Models with Minimal Scaling Cost CommFuse: Hiding Tail Latency via Communication Decomposition and Fusion for Distributed LLM Training A Taxonomy and Resolution Strategy for Client-Level Disagreements in Federated Learning Usable Agent Discovery for Decentralized AI Systems Cloud to Edge: Benchmarking LLM Inference On Hardware-Accelerated Single-Board Computers Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy Shard the Gradient, Scale the Model: Serverless Federated Aggregation via Gradient Partitioning Promoting Simple Agents: Ensemble Methods for Event-Log Prediction GraphLeap: Decoupling Graph Construction and Convolution for Vision GNN Acceleration on FPGA AGNT2: Autonomous Agent Economies on Interaction-Optimized Layer 2 Infrastructure FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels Stream-CQSA: Avoiding Out-of-Memory in Attention Computation via Flexible Workload Scheduling A Delta-Aware Orchestration Framework for Scalable Multi-Agent Edge Computing Federated Learning over Blockchain-Enabled Cloud Infrastructure Optimal Routing for Federated Learning over Dynamic Satellite Networks: Tractable or Not? Sherpa.ai Privacy-Preserving Multi-Party Entity Alignment without Intersection Disclosure for Noisy Identifiers
DiPETrans: A Framework for Distributed Parallel Execution...
Shrey Baheti, Parwat Singh Anjana, Sathya Peri, Yogesh Simmhan · 2019-06-27 · via cs.DC updates on arXiv.org

Contemporary blockchain such as Bitcoin and Ethereum execute transactions serially by miners and validators and determine the Proof-of-Work (PoW). Such serial execution is unable to exploit modern multi-core resources efficiently, hence limiting the system throughput and increasing the transaction acceptance latency. The objective of this work is to increase the transaction throughput by introducing parallel transaction execution using a static analysis technique. We propose a framework DiPETrans for the distributed execution of the transactions in a block. Here, peers in the blockchain network form a community to execute the transactions and find the PoW parallelly, using a leader-follower approach. During mining, the leader statically analyzes the transactions, creates different groups (shards) of independent transactions, and distributes them to followers to execute them in parallel. After the transaction executes, the community's compute power is utilized to solve the PoW concurrently. When a block is successfully created, the leader broadcasts the proposed block to other peers in the network for validation. On receiving a block, validators re-execute the block transactions and accept the block if they reach the same state as shared by the miner. Validation can also be done as a community, in parallel, following the same leader-follower approach as mining. We report experiments using over 5 Million real transactions from the Ethereum blockchain and execute them using our DiPETrans framework to empirically validate the benefits of our techniques over traditional sequential execution. We achieve a maximum speedup of 2.2x for the miner and 2.0x for the validator, with 100 to 500 transactions per block. Further, we achieve a peak of 5x end-to-end block creation speedup using a parallel miner over a serial miner when using 6 machines in the community.