Cover Image

Communication models for distributed Intel Xeon Phi coprocessors

The emergence of accelerator technology in current supercomputing systems is changing the landscape of supercom-puting architectures. Accelerators like GPGPUs and coprocessors are optimized for parallel computation while being more energy efficient. Their computational power per watt plays a crucial...

Full description

Saved in:
Bibliographic Details
Main Authors: Neuwirth, Sarah (Author) , Frey, Dirk (Author) , Brüning, Ulrich (Author)
Format: Chapter/Article Conference Paper
Language:English
Published: 2015
In: 2015 IEEE 21st International Conference on Parallel and Distributed Systems (ICPADS)
Year: 2015, Pages: 499-506
DOI:10.1109/ICPADS.2015.69
Subjects:
Konferenzschrift
Online Access:Resolving-System, Volltext: http://dx.doi.org/10.1109/ICPADS.2015.69
Verlag, Volltext: https://ieeexplore.ieee.org/document/7384332/
Get full text
Author Notes:Sarah Neuwirth, Dirk Frey and Ulrich Bruening

MARC

LEADER 00000caa a2200000 c 4500
001 1575433664
003 DE-627
005 20220814144146.0
007 cr uuu---uuuuu
008 180523s2015 xx |||||o 00| ||eng c
024 7 |a 10.1109/ICPADS.2015.69  |2 doi 
035 |a (DE-627)1575433664 
035 |a (DE-576)505433664 
035 |a (DE-599)BSZ505433664 
035 |a (OCoLC)1341009986 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 28  |2 sdnb 
100 1 |a Neuwirth, Sarah  |d 1986-  |e VerfasserIn  |0 (DE-588)1159979707  |0 (DE-627)1023096536  |0 (DE-576)505429004  |4 aut 
245 1 0 |a Communication models for distributed Intel Xeon Phi coprocessors  |c Sarah Neuwirth, Dirk Frey and Ulrich Bruening 
264 1 |c 2015 
300 |a 8 
336 |a Text  |b txt  |2 rdacontent 
337 |a Computermedien  |b c  |2 rdamedia 
338 |a Online-Ressource  |b cr  |2 rdacarrier 
500 |a Published online: 18 January 2016 
500 |a Gesehen am 23.05.2018 
520 |a The emergence of accelerator technology in current supercomputing systems is changing the landscape of supercom-puting architectures. Accelerators like GPGPUs and coprocessors are optimized for parallel computation while being more energy efficient. Their computational power per watt plays a crucial role in developing exaflop systems. Today's accelerators come with some limitations. They require a local host to configure and operate them. In addition, the number of host CPUs and accelerators does not scale independently. Another problem is the unbalanced communication between distributed accelerators. New communication frameworks are developed to optimize the internode communication. In this paper, four communication models using the Intel Xeon Phi coprocessor technology are compared. The Intel Xeon Phi coprocessor is based on the Intel Many Integrated Cores technology. It is an attractive accelerator due to its embedded Linux operating system, up to 1 TFLOPS of performance on a single chip, and its x86 64 compatibility. DCFA-MPI, MVAPICH2-MIC, and HAM-Offload are compared against the communication architecture for network-attached accelerators (NAA). Each communication model optimizes a different layer of the MIC communication architecture. The NAA approach makes the accelerator device independent from a local host system. Furthermore, it enables the accelerator to source and sink network traffic. Workloads can be dynamically assigned during run-time in an N to M ratio between CPUs and accelerators. The latency, bandwidth, and performance of the MPI communication layer of a prototype implementation are evaluated. 
650 4 |a application program interfaces 
650 4 |a Bismuth 
650 4 |a Computer architecture 
650 4 |a Coprocessors 
650 4 |a distributed accelerator 
650 4 |a distributed Intel Xeon Phi coprocessor 
650 4 |a HAM-Offload 
650 4 |a Intel many integrated core technology 
650 4 |a Intel Xeon Phi coprocessor technology 
650 4 |a internode communication 
650 4 |a Libraries 
650 4 |a Linux operating system 
650 4 |a MIC communication architecture 
650 4 |a Microwave integrated circuits 
650 4 |a MPI communication layer 
650 4 |a multiprocessing systems 
650 4 |a network-attached accelerator 
650 4 |a parallel computation 
650 4 |a parallel processing 
650 4 |a x86 64 compatibility 
655 7 |a Konferenzschrift  |0 (DE-588)1071861417  |0 (DE-627)826484824  |0 (DE-576)433375485  |2 gnd-content 
700 1 |a Frey, Dirk  |d 1985-  |e VerfasserIn  |0 (DE-588)1159982791  |0 (DE-627)1023100983  |0 (DE-576)505432978  |4 aut 
700 1 |a Brüning, Ulrich  |d 1954-  |e VerfasserIn  |0 (DE-588)1047086840  |0 (DE-627)777543397  |0 (DE-576)400390639  |4 aut 
773 0 8 |i Enthalten in  |a IEEE International Conference on Parallel and Distributed Systems (21. : 2015 : Melbourne)  |t 2015 IEEE 21st International Conference on Parallel and Distributed Systems (ICPADS)  |d Piscataway, NJ : IEEE, 2015  |g (2015), Seite 499-506  |h xxi, 858 Seiten  |w (DE-627)1656466198  |w (DE-576)50543136X  |z 9780769557854  |7 nnam 
773 1 8 |g year:2015  |g pages:499-506  |g extent:8  |a Communication models for distributed Intel Xeon Phi coprocessors 
856 4 0 |u http://dx.doi.org/10.1109/ICPADS.2015.69  |x Resolving-System  |x Verlag  |3 Volltext 
856 4 0 |u https://ieeexplore.ieee.org/document/7384332/  |x Verlag  |3 Volltext 
951 |a AR 
992 |a 20180523 
993 |a ConferencePaper 
994 |a 2015 
998 |g 1047086840  |a Brüning, Ulrich  |m 1047086840:Brüning, Ulrich  |d 700000  |d 720000  |e 700000PB1047086840  |e 720000PB1047086840  |k 0/700000/  |k 1/700000/720000/  |p 3  |y j 
998 |g 1159982791  |a Frey, Dirk  |m 1159982791:Frey, Dirk  |d 700000  |d 720000  |e 700000PF1159982791  |e 720000PF1159982791  |k 0/700000/  |k 1/700000/720000/  |p 2 
998 |g 1159979707  |a Neuwirth, Sarah  |m 1159979707:Neuwirth, Sarah  |d 700000  |d 720000  |e 700000PN1159979707  |e 720000PN1159979707  |k 0/700000/  |k 1/700000/720000/  |p 1  |x j 
999 |a KXP-PPN1575433664  |e 3009989423 
BIB |a Y 
JSO |a {"relHost":[{"id":{"isbn":["9780769557854"],"eki":["1656466198"]},"origin":[{"publisherPlace":"Piscataway, NJ","dateIssuedDisp":"2015","dateIssuedKey":"2015","publisher":"IEEE"}],"physDesc":[{"extent":"xxi, 858 Seiten"}],"title":[{"title_sort":"2015 IEEE 21st International Conference on Parallel and Distributed Systems (ICPADS)","title":"2015 IEEE 21st International Conference on Parallel and Distributed Systems (ICPADS)","subtitle":"proceedings : 14 - 17 December 2015, Melbourne, Victoria, Australia"}],"language":["eng"],"corporate":[{"roleDisplay":"VerfasserIn","display":"IEEE International Conference on Parallel and Distributed Systems (21., 2015, Melbourne)","role":"aut"},{"role":"isb","display":"Institute of Electrical and Electronics Engineers","roleDisplay":"Herausgebendes Organ"}],"recId":"1656466198","note":["Gesehen am 23.03.2018"],"disp":"IEEE International Conference on Parallel and Distributed Systems (21. : 2015 : Melbourne)2015 IEEE 21st International Conference on Parallel and Distributed Systems (ICPADS)","type":{"media":"Online-Ressource","bibl":"book"},"titleAlt":[{"title":"ICPADS 2015"},{"title":"21st International Conference on Parallel and Distributed Systems"}],"part":{"extent":"8","text":"(2015), Seite 499-506","pages":"499-506","year":"2015"}}],"physDesc":[{"extent":"8 S."}],"name":{"displayForm":["Sarah Neuwirth, Dirk Frey and Ulrich Bruening"]},"id":{"doi":["10.1109/ICPADS.2015.69"],"eki":["1575433664"]},"origin":[{"dateIssuedDisp":"2015","dateIssuedKey":"2015"}],"language":["eng"],"recId":"1575433664","type":{"bibl":"chapter","media":"Online-Ressource"},"note":["Published online: 18 January 2016","Gesehen am 23.05.2018"],"person":[{"role":"aut","roleDisplay":"VerfasserIn","display":"Neuwirth, Sarah","given":"Sarah","family":"Neuwirth"},{"role":"aut","display":"Frey, Dirk","roleDisplay":"VerfasserIn","given":"Dirk","family":"Frey"},{"roleDisplay":"VerfasserIn","display":"Brüning, Ulrich","role":"aut","family":"Brüning","given":"Ulrich"}],"title":[{"title_sort":"Communication models for distributed Intel Xeon Phi coprocessors","title":"Communication models for distributed Intel Xeon Phi coprocessors"}]} 
SRT |a NEUWIRTHSACOMMUNICAT2015 

Similar Items