In this paper, a 8×8 routing node has been designed using a high-performance digital cross-point switch with the throughput of 10Gbit/s. The communication latency is only 235 ns.
本文利用高性能数字交叉开关设计了8×8光互连路由结点,吞吐量为10Gbit/s,通信延迟仅为235 ns。
参考来源 - 计算机机群系统高速光互连路由结点及网络设计·2,447,543篇论文数据,部分数据来源于NoteExpress
Multithreads can not only implement fine-grain parallelism effectively, but also be helpful to hide communication latency with reasonable scheduling strategies.
多线程计算不仅用来实现细粒度的并行,合理的调度策略还有助于隐藏通信延迟。
By bypassing these steps and replacing them with direct guest kernel driver/colinux driver link, we can reduce the communication latency and thus enhance the cluster performance.
绕开这些步骤并将其替换 为直接的内核驱动程序/colinux 驱动程序连接,这样就可以减少通信延迟,从而提高集群的性能。
This avoids network latency and the overhead of inter-process communication, as well as context switching in the operating system, further contributing to extreme speed.
这样就能够避免网络延迟、进程间通信的开销以及操作系统中的上下文切换,从而进一步提高速度。
应用推荐