发表机构
NVIDIA Corporation(英伟达公司)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
针对LLM智能体任务修订,提出\retire版本化执行框架,通过撤销过时工作并认证可继承前缀,实现快速失效与选择性保留,减少首令牌时间中位数17.1%。
AI 中文摘要
当用户更改指令、工具失败或新信息改变计划时,LLM智能体会修订正在运行的任务。当今的服务器将修订表达为中止旧请求并提交替换请求。然而,旧执行的缓冲输出和未完成工作必须停止影响应用程序,而已完成的KV状态可能对其替换仍有用。分别处理这些义务可能会使过时效果可被发布,并迫使后继者重建有效状态。我们提出\retire,一种围绕\emph{版本化执行}的服务控制平面重新设计。请求拥有调度和内存资源;执行版本拥有\emph{权限},即发布输出或为当前执行安装状态的许可。\retire首先撤销过时工作,然后限制其剩余执行,并认证其后继者可继承的已完成前缀。后继者从该状态运行,同时异步回收隔离的旧资源。这在一个版本转换中统一了快速失效和选择性保留。我们在vLLM中跨输出发布、GPU执行、KV交接、分层恢复以及分布式和多租户服务实现了\retire。正确性实验验证了这些路径上的当前版本输出和有效状态继承。将失效与继承相结合,在受控配对实验中将修订到后继者首令牌时间的中位数减少了17.1%。对记录的编码智能体中断到达的重放未产生过时输出,并保持每个最终版本在重复修订中推进。\retire将中止并重启转变为协调交接,快速停止过时工作并为后继者保留有用工作。
英文摘要
LLM agents revise running tasks when users change instructions, tools fail, or new information changes a plan. Today's servers express a revision as aborting old requests and submitting replacements. Yet the old execution's buffered output and outstanding work must stop affecting the application, while completed KV state may still be useful to its replacement. Handling these obligations separately can leave obsolete effects publishable and force the successor to rebuild valid state. We present \retire, a serving control-plane redesign around \emph{versioned execution}. Requests own scheduling and memory resources; execution versions own \emph{authority}, the permission to publish output or install state for the current execution. \retire first revokes obsolete work, then bounds its remaining execution and certifies the completed prefix its successor can inherit. The successor runs from that state while isolated old resources are reclaimed asynchronously. This unifies fast invalidation and selective preservation in one version transition. We implement \retire in vLLM across output publication, GPU execution, KV handoff, tiered recovery, and distributed and multi-tenant serving. Correctness experiments verify current-version output and valid state inheritance across these paths. Combining invalidation with inheritance reduces revision-to-successor time-to-first-token by a median 17.1\% in controlled paired experiments. A replay of recorded coding-agent interruption arrivals emits no obsolete output and keeps every final version progressing through repeated revisions. \retire turns abort-and-restart into a coordinated handoff that stops obsolete work quickly and preserves useful work for its successor.