Disaggregated serving
Disaggregated serving is an LLM inference architecture that physically separates the prefill phase and the decode phase of text generation onto different sets of GPU hardware.
Artificial Intelligence
Explore Artificial Intelligence through related topics and the articles other pages reference most.
Explore articles
Browse subtopics: AI Inference
Articles that also belong to these categories. Counts cover all of Artificial Intelligence.
All articles (272)Machine Learning (54)Large Language Models (37)AI Tools & Products (31)AI Safety (24)Natural Language Processing (24)AI Companies (21)Robotics (21)AI Agents (19)Generative AI (17)Deep Learning (16)AI Ethics (14)Computer Vision (13)AI Hardware (10)AI History (10)Open Source AI (10)Reasoning Models (9)AI Benchmarks (8)AI Code Generation (8)Developer Tools (8)AI for Science (7)Computer Science (7)Google (7)Reinforcement Learning (7)AI Policy & Regulation (6)Enterprise AI (6)People (6)AI Infrastructure (5)AI Research (5)Anthropic (5)OpenAI (5)AI in Gaming (4)Autonomous Vehicles (4)Data Science (4)Google DeepMind (4)Humanoid Robots (4)Information Retrieval (4)Multimodal AI (4)Neural Networks (4)Prompt Engineering (4)Software Development (4)Speech & Audio AI (4)AI Models (3)Embodied AI (3)Training & Optimization (3)Video Generation (3)AI Events (2)AI Inference (2)Algorithms (2)Chinese AI (2)Conversational AI (2)Healthcare AI (2)Image Generation (2)Research Organizations (2)Statistics (2)2025 in artificial intelligence (1)Agentic Commerce (1)ChatGPT (1)China AI (1)Data & Datasets (1)Drug Discovery (1)Finance AI (1)Mathematics (1)Meta AI (1)Microsoft (1)Model Architecture (1)Model Evaluation (1)Programming Languages (1)Transformer Models (1)Venture Capital (1)World Models (1)
Showing 1-2 of 2 articles
Inference-time scaling
Inference-time scaling (also called test-time compute scaling) is the practice of improving an AI model's output quality by allocating more computational resources during inference rather than during training.