Groq 揭秘 LPU 架构：如何无损加速万亿参数模型推理
原标题：Moonshot’s Kimi K2 recently launched in preview on GroqCloud and developers keep asking us: how is Groq running a 1-trillion-parameter model this fast?
AI 摘要
Groq 在 GroqCloud 上预览了 Moonshot 的 Kimi K2 模型，并解释了其 LPU 架构如何通过 TruePoint 数值格式、SRAM 存储和静态调度等技术，在不损失精度的情况下实现万亿参数模型的快速推理。该架构通过选择性降低精度和优化内存层次，实现了比传统 GPU 更高的推理速度和更低的延迟。
正文节选
Moonshot’s Kimi K2 recently launched in preview on GroqCloud and developers keep asking us: how is Groq running a 1-trillion-parameter model this fast? Legacy hardware forces a choice: faster inference with quality degradation, or accurate inference with unacceptable latency. This tradeoff exists because GPU architectures optimize for training workloads. The LPU–purpose-built hardware for inference–preserves quality while eliminating architectural bottlenecks which create latency in the first pl