Influence of the direct message search mechanism based on the TCP protocols on the exchange process

作者: Vasyl Melnyk , Kateryna Melnyk , Svitlana Lavrenchuk , Ihor Burchak , Oleksii Kaganiuk

DOI: 10.15587/1729-4061.2019.167995

关键词:

摘要: The direct search mechanism is implemented with the expansion of traditional socket TCP interface for receiving messages while bypassing order established queue. This can be used high-performance and clustered computer systems in to intensify data exchange continuous support a maximum load on computing machines. message base Linux kernel. experimental test results are obtained by using set simple microbenchmarks. During test, sender sends required number fixed-size via an connection, receiver skips unexpected reads expected one into user's space. approach finding realized multiple searches case where application treats as list ability receive delete not only from buffer top but any place it. All recognized processed developed seek_recv() call. Each contains ~80 repetitions, which include such operations opening, sending 800–1000 according acceptance politics, closure. system uses active at same time. received confirm noticeable decrease processing CPU time 36–40 % overall productivity growth. However, when volume approaching 1000 bytes, i.e. close typical size packet useful load, there process.

参考文章(13)
Torsten Hoefler, Torsten Mehlan, Andrew Lumsdaine, Wolfgang Rehm, Netgauge: a network performance measurement framework high performance computing and communications. pp. 659- 671 ,(2007) , 10.1007/978-3-540-75444-2_62
Thomas F.J.M. Pasquier, Jatinder Singh, Jean Bacon, Olivier Hermant, Managing Big Data with Information Flow Control international conference on cloud computing. pp. 524- 531 ,(2015) , 10.1109/CLOUD.2015.76
V. Paxson, Q. Xie, L. Zhang, C. Sharp, M. Kalla, T. Taylor, R. Stewart, I. Rytina, H. Schwarzbauer, K. Morneault, Stream Control Transmission Protocol RFC. ,vol. 2960, pp. 1- 134 ,(2000)
Катерина Вікторівна Мельник Василь Михайлович Мельник, Наталія Володимирівна Багнюк, None, Вплив високопродуктивних сокетів на інтенсивність обробки даних ScienceRise. ,vol. 6, pp. 38- 48 ,(2015) , 10.15587/2313-8416.2015.44380
Semra Aydin, Omer Faruk Bay, Building a high performance computing clusters to use in computing course applications Procedia - Social and Behavioral Sciences. ,vol. 1, pp. 2396- 2401 ,(2009) , 10.1016/J.SBSPRO.2009.01.420
Iain Bethune, J Mark Bull, Nicholas J Dingle, Nicholas J Higham, Performance analysis of asynchronous Jacobi's method implemented in MPI, SHMEM and OpenMP ieee international conference on high performance computing data and analytics. ,vol. 28, pp. 97- 111 ,(2014) , 10.1177/1094342013493123
Michael D Beynon, Tahsin Kurc, Umit Catalyurek, Chialin Chang, Alan Sussman, Joel Saltz, Distributed processing of very large datasets with DataCutter parallel computing. ,vol. 27, pp. 1457- 1478 ,(2001) , 10.1016/S0167-8191(01)00099-0
I. Pratt, K. Fraser, Arsenic: a user-accessible gigabit Ethernet interface international conference on computer communications. ,vol. 1, pp. 67- 76 ,(2001) , 10.1109/INFCOM.2001.916688
Philip Carns, Sam Lang, Robert Ross, Murali Vilayannur, Julian Kunkel, Thomas Ludwig, Small-file access in parallel file systems international parallel and distributed processing symposium. pp. 1- 11 ,(2009) , 10.1109/IPDPS.2009.5161029