送交者: Nixrreg 于 2011-02-24, 00:22:12:

回答: 今天非常成功，把system 其中的一个major component 由 steven 于 2011-02-24, 00:01:24:

• the goal is always constant latency, logN is the second best. (无内容) - steven (0 bytes) 2011-02-24, 01:40:41 (551902)
  • how can you measure latency with big O notation in an concurrent environment? (无内容) - ak70 (0 bytes) 2011-02-24, 02:10:43 (551904)
    • why not? (无内容) - steven (0 bytes) 2011-02-24, 02:26:17 (551906)
      • because of the concurrency issues. Unless the algorithm - ak70 (255 bytes) 2011-02-24, 02:33:08 (551909)
        • big O works for concurrent also, - steven (109 bytes) 2011-02-24, 02:50:15 (551913)
          • what about the delays caused by resource locking? - ak70 (139 bytes) 2011-02-24, 03:19:35 (551915)
            • The delay for syncrhonizing tasks is no different than any other delay. - steven (34 bytes) 2011-02-24, 10:32:40 (552047)
• 减低latency又不是算法优化，用不着O之类的东西。 (无内容) - ak70 (0 bytes) 2011-02-24, 00:43:59 (551888)
  • broadcast不是么？ (无内容) - Nixrreg (0 bytes) 2011-02-24, 00:48:12 (551890)
    • steven是做编程的吧？他说的应该是系统底层的东西。 (无内容) - ak70 (0 bytes) 2011-02-24, 00:52:30 (551894)

加跟贴