Google TPU v6e (Trillium)
芯片
AI加速器
发货中
The sixth generation: ~4.7x v5e compute, used to train Gemini 2.0.
您正在查看一条芯片记录:该处理器是什么、其关键规格、谁设计它、谁制造它、谁供应其零部件,以及它被用在哪里。
Trillium raised per-chip compute about 4.7x over v5e with doubled HBM capacity and bandwidth, and Google stated Gemini 2.0 was trained on it. It became generally available in December 2024 and anchors Google's inference fleet alongside Ironwood.
规格与关键参数
每个数据均标注来源依据:厂商规格、厂商声明、公开报道、估算或官方数据。"未公开披露"即字面含义——我们从不捏造数值。
Overview
Designer
Google
Announced
May 2024; GA December 2024
内存
Memory
32 GB HBM, 1.6 TB/sVendor spec
算力
BF16
918 TFLOPSVendor spec
Improvement
~4.7x v5e peak compute per chipVendor claim
Physical
Pod scale
256 chips per pod, multi-pod via Jupiter fabricVendor spec
它如何与整个技术栈相连
本记录的关联链接一览——点击分组可直接跳转。
前代产品
来源
记录审阅截至 2026-01-15 — 查看 方法论.