问 HN：Lichess 和 Stockfish 之间的差异
7 分•作者: HNLurker2•6 个月前
我正在试图理解 Lichess 分析棋盘和我自己的 Stockfish 设置之间的一个差异。
在 Lichess（基于浏览器的分析）上，Stockfish 在我的 Redmi Note 14 Pro 上报告接近 1 MN/s。然而，当我通过 Python 程序使用原生可执行文件在本地运行 Stockfish 时，我只看到大约 600 kN/s。
令人困惑的是，尽管报告的速度更高，但 Lichess 需要大约 2 分 30 秒才能达到深度 30，而我的本地设置大约在 53 秒内达到深度 30，即使它报告的 N/s 较低。Lichess 在频繁的评估更新方面也显得更加“活跃”。
我怀疑这与 N/s 的测量或显示方式（瞬时 vs 平均）、搜索配置的差异（连续搜索 vs 重启、MultiPV、哈希复用），或者引擎驱动方式（例如，UI 或 I/O 限制）有关。这也引出了一个问题，即“深度 30”在不同的前端之间是否可以直接比较。
有人研究过 Lichess 如何报告 Stockfish 速度，或者为什么显示更高 N/s 的设置仍然需要更长的时间才能达到相同的标称深度吗？
查看原文
I’m trying to understand a discrepancy between Lichess’s analysis board and my own Stockfish setup.<p>On Lichess (browser-based analysis), Stockfish reports close to 1 MN/s on my Redmi Note 14 Pro. However, when I run Stockfish locally via a Python program using the native executable, I only see around 600 kN/s.<p>What’s confusing is that despite the higher reported speed, Lichess takes about 2:30 to reach depth 30, while my local setup reaches depth 30 in about 53 seconds, even though it reports a lower N/s. Lichess also appears much more “active” in terms of frequent evaluation updates.<p>I suspect this has to do with how N/s is measured or displayed (instantaneous vs average), differences in search configuration (continuous search vs restarts, MultiPV, hash reuse), or overhead from the way the engine is driven (e.g., UI or I/O throttling). It also raises the question of whether “depth 30” is directly comparable across different frontends.<p>Has anyone looked into how Lichess reports Stockfish speed, or why a setup showing higher N/s can still take significantly longer to reach the same nominal depth?