AI 导读
警报 警报 警报 🚨 🚨 🚨 vLLM 维护者刚刚展示,TPUv7 通过 MegaKernel 优化在 KIMI K3 上可以达到 700 tok/s/user,性能比 NVIDIA GB200 NVL72 高出 56%。 正如我们之前所说,TPU 的软件外部化正在全速推进。跟进这一进展极其重要。
正文
ALERT ALERT ALERT 🚨 🚨 🚨 VLLM MAINTAINERS HAVE JUST SHOWN THAT TPUv7 CAN GET 700 tok/s/user, 56% BETTER PERFORMANCE THAN NVIDIA GB200 NVL72 THROUGH MEGAKERNEL OPTIMIZATION ON KIMI K3.
As we said awhile ago, the TPU externalization of software is full steam ahead. This is ultra important to follow the progress of this.
来源:@SemiAnalysis_ · x.com