← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6485.g · 生成模型:GPT-6

规模运行时间不证明加速

理解等级:高级

固定配置下记录不同数据量的运行时间,可以描述负载变化,但不能单独证明并行加速。比较加速还需相同任务的基线、硬件、工作节点与重复测量。本章图5没有给出充分对照,不能据此量化MapReduce相对优势。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      frequency value and is stored in a format we can use later by unpacking the key. Then, we can load the data, unpack it and search the author's name to get his/her specialization. In this approach, the computation must be done only once because of parallel execution. In Fig. (5), the graph shows the parallel programming nature and time consumption with the increased data size.

    参考来源

    - 位置:《Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications》第05章《主题建模与作者专长推荐》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications/_epub-src/OEBPS/Text/9789815136746-C5.html) - 结论依据:摘录中关于规模运行时间不证明加速的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:frequency value and is stored in a format we can use later by unpacking the key. Then, we can load the data, unpack it and search the author's name to get his/her specialization. In this approach, the computation must be done only once because of parallel execution. In Fig. (5), the graph shows the parallel programming nature and time consumption with the increased data size.