A High-Performance Network Interface Architecture for NoCs Using Reorder Buffer Sharing

Author(s):  
Masoumeh Ebrahimi ◽  
Masoud Daneshtalab ◽  
Pasi Liljeberg ◽  
Juha Plosila ◽  
Hannu Tenhunen
2001 ◽  
Vol 9 (4) ◽  
pp. 211-222 ◽  
Author(s):  
Marian Bubak ◽  
Dariusz Żbik ◽  
Dick van Albada ◽  
Kamil Iskra ◽  
Peter Sloot

Efficient load balancing is essential for parallel distributed computing. Many parallel computing environments use TCP or UDP through the socket interface as a communication mechanism. This paper presents the design and development of a prototype implementation of a network interface that can preserve communication between processes during process migration. This new communication library is a substitution for the well-known socket interface. It is implemented in user — space; it is portable, and no modifications of user applications are required. TCP/IP is applied for internal communication, which guarantees relatively high performance and portability.


2018 ◽  
Vol 76 ◽  
pp. 69-80
Author(s):  
Seyed Mohamad Taghi Adl ◽  
Siamak Mohammadi

2014 ◽  
Vol 543-547 ◽  
pp. 3118-3121 ◽  
Author(s):  
Ya Ming Ke ◽  
Yong Bin Wang ◽  
Ying Li

With the development of computer technology, the performance of hardware is no longer the primary problem that influences the performance of the web server. Just improving the hardware processing speed of the web server is invalid in practical applications, so the performance of the web server should be improved in the software layer. The improving methods focus on the followings: In the aspect of the network IO, the EPOLL structure is used. The EPOLL structure can not only deal with mass network connections, but also have a better performance for processing network data. Those improvements all enhance the throughput of the system effectively. In the aspect of the hard disk IO, the asynchronous IO is used to replace the synchronous blocking IO, so the hard disk IO and other requisitions can be processed simultaneously. Besides, the hard disk IO requisitions are processed by kernel instead of application layer, so the number of concurrent request and the usage effectiveness of CPU can be improved at the same time. This means the processing capacity of the web server can be improved consequently. At last, aiming at moving data effectively from file to the network interface, the zero-copy technology is used to optimize especially. This helps the improvement of the web servers business processing capacity a lot. Above methods are used to put forward a high performance web server model which is based on the asynchronous IO and the zero-copy technology. The web server model is realized at the foundation of the analysis of several traditional web server models. Those lay the groundwork for the further study of a high performance web server model which is more practical and has more abundant functions.


2007 ◽  
Vol 18 (9) ◽  
pp. 1282-1295 ◽  
Author(s):  
Konosuke Watanabe ◽  
Tomohiro Otsuka ◽  
Junichiro Tsuchiya ◽  
Hiroaki Nishi ◽  
Junji Yamamoto ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document