今日速览
| 序号 | 标题 | 来源 | 日期 | 主题 | 推荐等级 |
|---|---|---|---|---|---|
| 1 | AutoPass: Evidence-Guided LLM Agents for Compiler Performance Tuning | arXiv | 2026-06-18 | AI-Agent | 高 |
| 2 | PACMS: Submodular Context Selection as a Pluggable Engine for LLM Agents | arXiv | 2026-06-18 | AI-Agent | 高 |
重点论文与技术动态
1. AutoPass: Evidence-Guided LLM Agents for Compiler Performance Tuning
- 来源：arXiv
- 日期：2026-06-18
- 作者/机构：Zepeng Li, Jie Ren, Zhanyong Tang, Jie Zheng, Zheng Wang
- 主题标签：
AI-Agent,arXiv - 推荐等级：高
- 分类：cs.SE, cs.AI
一句话结论
AutoPass通过证据引导的LLM代理框架实现编译器性能调优，在x86-64和ARM64系统上分别实现1.043倍和1.117倍的几何平均加速，显著优于专家调优和传统自动调优方法。
核心内容
- AutoPass开放编译器内部状态，允许LLM查询优化状态并分析中间表示，避免将编译器视为黑盒的传统局限。
- 采用迭代搜索过程，利用运行时反馈诊断性能回归并指导延迟优化决策，提升调优精度。
- 无需离线训练或任务特定微调，可直接应用于新基准和平台，实现快速部署。
方法与数据
- 方法：基于LLM的多代理框架，结合编译器和运行时证据驱动优化决策。
- 摘要未明确。
价值判断
- 值得关注：创新性地解决复杂微架构效应和噪声测量问题，将LLM深度融入编译器内部交互。
- 可复用点：推理-only设置无需额外训练，便于迁移至不同系统和基准测试。
- 局限/待核查：摘要未明确具体局限，需验证在更广泛硬件场景的适用性。
摘要
Large Language Models (LLMs) show promise for code compilation tasks, but applying them to runtime performance tuning is difficult due to complex microarchitectural effects and noisy runtime measurements. We present AutoPass, a multi-agent framework for compiler performance tuning that uses compiler and runtime evidence to guide LLM-generated optimization decisions. Rather than treating the compiler as a black box like prior auto-tuning schemes, AutoPass opens up the compiler to the LLM, enabling it to query compiler-internal optimization states and analyze the intermediate representation to orchestrate compiler options. The search process iteratively refines optimization configurations using measured runtime feedback to diagnose regressions and guide latency-improving edits. AutoPass operates in an inference-only, training-free setting and requires no offline training or task-specific fine-tuning, making it readily applicable to new benchmarks and platforms. We implement AutoPass on the LLVM compiler and evaluate it on server-grade x86-64 and embedded ARM64 systems. AutoPass outperforms expert-tuned heuristics and classical autotuning methods, achieving geometric-mean speedups of 1.043x and 1.117x over LLVM -O3 on x86-64 and ARM64, respectively.2. PACMS: Submodular Context Selection as a Pluggable Engine for LLM Agents
- 来源：arXiv
- 日期：2026-06-18
- 作者/机构：Manu Ghulyani, Arunabh Singh, Karan Bharadwaj, Ankit Nath, Suranjan Goswami
- 主题标签：
AI-Agent,arXiv - 推荐等级：高
- 分类：cs.IR
一句话结论
PACMS提出了一种基于子模优化的上下文选择机制，作为LLM代理的可插拔引擎，能有效解决上下文窗口过载问题，避免传统按时间顺序截断导致的关键信息丢失。
核心内容
- LLM代理上下文包含对话回合、记忆条目和工具输出，超限时现有按时间截断方法主题盲，导致早期重要事实被丢弃（如当前查询涉及该事实时），而无关近期内容被保留。
- 现有方案如检索增强生成无法处理代理已有的上下文池，上下文压缩则操作查询盲且有损，均未将记忆、对话和工具输出统一管理。
- PACMS将所有上下文元素视为单一候选池，通过子模优化在提示组装时基于相关性选择保留内容，显著提升跨多轮对话的记忆效率。
方法与数据
- 摘要未明确具体方法实现和实验数据。
价值判断
- 值得关注：针对LLM代理中上下文管理的核心痛点，尤其在需要跨多轮对话回忆信息的场景，解决传统方法失效问题。
- 可复用点：子模优化框架可作为可插拔组件无缝集成到现有LLM代理架构，无需重构核心流程。
- 局限/待核查：摘要未明确提及具体局限性或实验验证细节。
摘要
Conversational and tool-using LLM agents operate over a context window that fills from several directions simultaneously. As a session proceeds, the agent accumulates user and assistant turns, entries drawn from a persistent memory store, and often largest of all, the verbatim outputs of tool calls such as file reads, search results, and API responses. Once the cumulative context exceeds the model's token budget, the framework must decide what to keep. The prevailing mechanism is recency truncation, sometimes paired with periodic summarization. This is topic-blind: a fact established early in a session is discarded simply because it is old, even when the current user query is about exactly that fact; conversely, verbose but irrelevant recent material is retained. Agents that must recall information across many turns, the defining case for memory, are precisely where recency truncation fails. Existing alternatives sit outside the agent's assembly step. Retrieval augmented generation fetches external documents into the prompt but does not arbitrate the agent's \emph{already-present} pooled context. Context-compression methods reduce token count by rewriting or pruning text, but operate query-blind and lossily. Neither treats memory entries, conversation turns, and tool outputs as a single candidate pool to be selected from by relevance at the moment the prompt is assembled.-
AutoPass: Evidence-Guided LLM Agents for Compiler Performance Tuning AutoPass通过证据引导的LLM代理框架实现编译器性能调优，在x86-64和ARM64系统上分别实现1.043倍和1.117倍的几何平均加速，显著优于专家调优和传统自动调优方法。
-
PACMS: Submodular Context Selection as a Pluggable Engine for LLM Agents PACMS提出了一种基于子模优化的上下文选择机制，作为LLM代理的可插拔引擎，能有效解决上下文窗口过载问题，避免传统按时间顺序截断导致的关键信息丢失。