架构中粗粒度/细粒度的定义
在谈论 API 时，经常使用术语“粗粒度”或“细粒度”。这些是什么意思/有什么例子吗？
谢谢
When talking about APIs, the terms "coarsely grained" or "finely grained" are used a lot. What do these mean/are there any examples?
Thanks
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
通常指 FPGA（现场可编程门阵列）的最小逻辑元件。例如，如果它是 NAND 门，则它被认为是细粒度的。如果它是一个能够与许多 NAND 等效的复杂逻辑块，那么它就是粗粒度的。很容易看出，如果综合器选择一个复杂的块来执行“几个”门的逻辑，则会浪费芯片面积和利用率，最好留给细粒度架构。如果您想使用“大量与非门”进行浮点数学计算，那么效率不如具有浮点的粗粒度 FPGA，甚至是整数、加法器、桶式移位器和/或 PLA 元件，可以是级联更大的位字。这将构成一个快速加法器，并且占用的面积比细粒度的加法器要少，并且可以接近 100% 地利用复杂逻辑块。
Often in reference to FPGA's (Field Programmable Gate Arrays) smallest logic element. If it's, for example, a NAND gate, it is considered fine grained. If it's a complex logic block capable of MANY NAND equivalences, it is coarse grained. It is easy to see that if the synthesizer picks a complex block to do a "few" gates worth of logic, it was a waste of the chip area and utilization and best left for fine grained architectures. If you wanted to do floating point math with a "sea of nand gates" that would not be as efficient as a coarse grained FPGA that has floating point, or even integer, adders, barrel shifters, and or PLA elements in them that can be cascaded for larger bit words. that would make a fast adder and take less area than fine grained and utilize closer to 100% of the complex logic block.
我已经看到 SOA 中经常使用这种区别，例如：
http
基本上，粗粒度系统比细粒度系统具有更少的方法调用。显然，与细分的细粒度版本相比，粗粒度方法的数量更少，工作量更多。
当调用成本高昂时（例如在分布式系统中），粗粒度接口通常会更好。
I've seen the distinction used a lot with SOA, e.g.:
http://www.ibm.com/developerworks/webservices/library/ws-soa-granularity/
Basically, coarse-grained systems will have less method calls than a fine-grained system. Clearly coarse-grained methods will be fewer and do more work than the broken-down fine-grained versions.
Coarse grained interfaces will generally be better where calls are expensive, in a distributed system for example.