Event ·Multi-source

Together Compiler 发布:智能体自动为模型和硬件优化内核

Together 推出 Together Compiler,用智能体针对用户的模型、硬件和流量持续优化 GPU 内核。该系统源自 FlashAttention、ThunderKittens、ThunderAgent 背后团队的成果。在其研究员已高度手动优化的 serverless 端点上,中位 TPS 提升 12% 到 20%。已有生产环境客户将响应与生成时间缩短约三分之一到近一半。

Current read

Together 出了个自动优化 GPU 内核的智能体,在你已调优过的端点上还能把速度再提 12%-20%,生产响应时间砍近一半。

3 sources64° AI heatLast updated 2026/11/6 22:56:00

Evidence

3 sources
Related source 1Tri Dao (FlashAttention)
View source
Related source 2marktechpost
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.