Primary sourcearXiv cs.AI
View sourceEvent ·Multi-source
Evolving Agent Teams:多智能体框架在 GPU 内核优化上超越 FlashInfer
多智能体框架 Evolving Agent Teams 针对现有编程智能体在专家级内核上陷入局部最优的问题,用中央监督者诊断并重组智能体团队,配合包含学习文档和演化历史的持久文件系统积累经验。在相同 token 预算下,该框架持续跑赢主流编程智能体基线。扩大预算后,它从零写出的 CUDA 内核在 FlashInfer-Bench 47 个生产负载中赢下 36 个,平均加速达到 FlashInfer MLA paged decode 内核的 1.2 倍。其他任务上也有具体成绩:AMD MI350X 上 KDA prefill 相比 vLLM 达 3.01 倍,NVIDIA B200 上相比 FlashKDA 达 3.44 倍,Apple M4 Pro 上 MoE 专家卸载服务达到 30.3 tokens/s 解码速度。
Current read
一个多智能体框架从零写出了比 FlashInfer 还快 1.2 倍的 CUDA 内核,还附了开源代码,做内核优化的可以试试。
8 sources58° AI heatLast updated 2026/10/27 08:46:21
Evidence
8 sourcesRelated source 1面壁智能 ModelBest
View sourceRelated source 2rohanpaul_ai
View sourceRelated source 3NVIDIA
View sourceRelated source 4Artificial Analysis
View sourceRelated source 5Tech in Asia
View sourceRelated source 6Decoder
View sourceRelated source 7Genk
View sourceConflict 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.