Unverified60% confidenceFactExact time

Large model inference is divided into two phases, Prefill and Decode; Prefill is a major component of Time to First Token (TTFT), while Decode corresponds to Time Per Output Token (TPOT).

2

Sources

60%

Confidence

Long-term

Relevance

9/11/2026

First Seen

Sources

Prompt Cache Deep Dive: From KV Cache to Agent Engineering Practice

Related Entities

Prefillconcept Decodeconcept TTFTconcept TPOTconcept

Related Claims

VerifiedLLM 推理分为预填充阶段（Prefill，计算密集型）和解码阶段（Decode，内存带宽密集型）两个阶段76% similar UnverifiedLLM推理的prefill阶段并行处理输入所有token并生成初始KV Cache，是计算密集型操作，是TTFT的直接决定因素75% similar Verified推测解码（Speculative Decoding）是目前最主流的大模型推理加速路径之一，使用小型草稿模型快速生成候选token再由主模型并行验证68% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/908129

API

curl https://kongchang.com/api/v1/knowledge/claims/908129

MCP

get_claim(id=908129)