Low-latency and zero-copy message passing protocols for SCI-based clusters
Resumo
This paper presents the design and implementation of three message passing protocols, whose development aimed at efficiently exploiting the high-performance capabilities of the SCI interconnect. These protocols are compared to the communication mechanisms adopted by MPI implementations for SCI clusters. The performance of the proposed protocols allows us to state that they are more efficient, in terms of latency and bandwidth, than the correspondent communication strategies employed by existing MPI implementations specifically designed for SCI-connected clusters.
Palavras-chave:
message passing, high-performance networks, cluster computing, SCI
Referências
BARRETO, M. et al. lmplementation of the DECK environment with BIP. In: MYRINET USER GROUP CONFERENCE, 1., 2000. Lyon. France. Proceedings... Lyon: INRIA Rocquencourt, 2000. p.82-88.
BODE,. N. et al. Myrinet: a gigabit-per-second local-area network. IEEE Micro, Los Alamitos, v. 15, n. 1, p.29-36, Feb. 1995.
BUYYA, Rajkumar (Ed.). High performance cluster computing: architectures and systems. Upper Saddle River: Prentice Hall PTR, 1999. 849p.
FISCHER, M.; SIMON, J. Embedding SCI into PVM. In: EUROPEAN PVM/MPI USERS GROUP MEETING, 4 .. 1997, Cracow. Proceedings ... Berlin: Springer-Verlag, 1997. p. 177-184. (Lecture Notes in Computer Science, v. l332).
FISCHER, Markus; REINEFELD, Alexander. PVM for SCI clusters. In: HELLWAGNER. Hermann: REINEFELD. Alexander (Eds.). SCI: Scalable Coherent Interface: architecture and software for high-performance compute clusters. Berlin: Springer, 1999. p.239-248. (Lecture Notes in Computer Science, v.1734).
GEIST, AI et al. PVM: parallel virtual machine. Cambridge. USA: MIT Press. 1994.
GIACOMINI, F. et al. Low-level SCI software requirements, analysis and predesign. [S.I.]: ESPRIT Project 23174 - Software Infrastructure for SCI (SISCI), 1998.
GILLET, R. Memory Channel Network for PCI. IEEE Micro, v. 16, n. 1. p.12-18, Feb. 1996.
HELLWAGNER, Hermann; REINEFELD, Alexander (Eds.). SCI: Scalable Coherent Interface: architecture and software for high-performance compute clusters. Berlin: Springer. 1999. 490p. (Lecture Notes in Computer Science, v. 1734).
HERLAND, B. G.; EBERL, M.; HELLWAGNER, H. A common messaging layer for MPI and PVM over SCI. In: HIGH PERFORMANCE COMPUTING AND NETWORKING, 1998. Amsterdam. Proceedings... Berlin: Springer-Verlag. 1998. p.576-587. (Lecture Notes in Computer Science, v. 1401).
HUSE, L. P. et al. ScaMPI-design and implementation. In: HELLWAGNER. Hermann: REINEFELD, Alexander (Eds.). SCI: Scalable Coherent Interface: architecture and software for high-performance compute clusters. Berlin: Springer. 1999. p.249-261. (Lecture Notes in Computer Science, v. 1734).
INSTITUTE OF ELECTRICAL AND ELECTRONIC ENGINEERS. IEEE standard for scalable coherent interface (SCI). New York: [s.n.]. 1992. IEEE 1596-1992.
MPI FORUM. The MPI message passing interface standard. Knoxville: University of Tennessee. 1994.
OLIVEIRA, Fábio Abreu Dias de. Uma biblioteca para programação paralela por troca de mensagens de clusters baseados na tecnologia SCI. Porto Alegre: PPGC da UFRGS, 2001. Master's Thesis. (In Portuguese).
OMANG. Knut. Synchronization support in I/O adapter based SCI. In: INTERNATIONAL WORKSHOP ON COMMUNICATION AND ARCHITECTURAL SUPPORT FOR NETWORK-BASED PARALLEL COMPUTING. 1.. 1997. San Antonio. Texas. Proceedings... Berlin: Springer-Verlag. 1997. p. 158-172. (Lecture Notes in Computer Science, v. 1199).
RYAN, Stein Jorgen; GJESSING, Stein; LIAAEN, Marius. Cluster communication using a PCI to SCI interface. In: INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED COMPUTING AND SYSTEMS. 8 .. 1996. Chicago. Proceedings . . . [S. I.: s.n.]. 1996.
WORRINGEN, Joachim: BEMMERL. Thomas. MPICH for SCI-connected clusters. In: SCI-EUROPE. 1999, Toulouse. France. Proccedings ... [S.I.: s.n.]. 1999. p.3-11.
WORRINGEN, Joachirn. SCI-MPICH: the second generation. In: SCI EUROPE, 3., 2000, Munich, Germany. Proceedings ... [S.I.: s.n.], 2000. p. 10-20. Organizado como conference stream do Euro-Par'2000.
ZORAJA, Ivan; HELLWAGNER, Herrmann; SUNDERAM. Vaidy. SCIPVM: parallel distributed cornputing on SCI workstation clusters. Concurrency: Practice and Experience. v.II. n.13, p. 121-138, Mar. 1999.
BODE,. N. et al. Myrinet: a gigabit-per-second local-area network. IEEE Micro, Los Alamitos, v. 15, n. 1, p.29-36, Feb. 1995.
BUYYA, Rajkumar (Ed.). High performance cluster computing: architectures and systems. Upper Saddle River: Prentice Hall PTR, 1999. 849p.
FISCHER, M.; SIMON, J. Embedding SCI into PVM. In: EUROPEAN PVM/MPI USERS GROUP MEETING, 4 .. 1997, Cracow. Proceedings ... Berlin: Springer-Verlag, 1997. p. 177-184. (Lecture Notes in Computer Science, v. l332).
FISCHER, Markus; REINEFELD, Alexander. PVM for SCI clusters. In: HELLWAGNER. Hermann: REINEFELD. Alexander (Eds.). SCI: Scalable Coherent Interface: architecture and software for high-performance compute clusters. Berlin: Springer, 1999. p.239-248. (Lecture Notes in Computer Science, v.1734).
GEIST, AI et al. PVM: parallel virtual machine. Cambridge. USA: MIT Press. 1994.
GIACOMINI, F. et al. Low-level SCI software requirements, analysis and predesign. [S.I.]: ESPRIT Project 23174 - Software Infrastructure for SCI (SISCI), 1998.
GILLET, R. Memory Channel Network for PCI. IEEE Micro, v. 16, n. 1. p.12-18, Feb. 1996.
HELLWAGNER, Hermann; REINEFELD, Alexander (Eds.). SCI: Scalable Coherent Interface: architecture and software for high-performance compute clusters. Berlin: Springer. 1999. 490p. (Lecture Notes in Computer Science, v. 1734).
HERLAND, B. G.; EBERL, M.; HELLWAGNER, H. A common messaging layer for MPI and PVM over SCI. In: HIGH PERFORMANCE COMPUTING AND NETWORKING, 1998. Amsterdam. Proceedings... Berlin: Springer-Verlag. 1998. p.576-587. (Lecture Notes in Computer Science, v. 1401).
HUSE, L. P. et al. ScaMPI-design and implementation. In: HELLWAGNER. Hermann: REINEFELD, Alexander (Eds.). SCI: Scalable Coherent Interface: architecture and software for high-performance compute clusters. Berlin: Springer. 1999. p.249-261. (Lecture Notes in Computer Science, v. 1734).
INSTITUTE OF ELECTRICAL AND ELECTRONIC ENGINEERS. IEEE standard for scalable coherent interface (SCI). New York: [s.n.]. 1992. IEEE 1596-1992.
MPI FORUM. The MPI message passing interface standard. Knoxville: University of Tennessee. 1994.
OLIVEIRA, Fábio Abreu Dias de. Uma biblioteca para programação paralela por troca de mensagens de clusters baseados na tecnologia SCI. Porto Alegre: PPGC da UFRGS, 2001. Master's Thesis. (In Portuguese).
OMANG. Knut. Synchronization support in I/O adapter based SCI. In: INTERNATIONAL WORKSHOP ON COMMUNICATION AND ARCHITECTURAL SUPPORT FOR NETWORK-BASED PARALLEL COMPUTING. 1.. 1997. San Antonio. Texas. Proceedings... Berlin: Springer-Verlag. 1997. p. 158-172. (Lecture Notes in Computer Science, v. 1199).
RYAN, Stein Jorgen; GJESSING, Stein; LIAAEN, Marius. Cluster communication using a PCI to SCI interface. In: INTERNATIONAL CONFERENCE ON PARALLEL AND DISTRIBUTED COMPUTING AND SYSTEMS. 8 .. 1996. Chicago. Proceedings . . . [S. I.: s.n.]. 1996.
WORRINGEN, Joachim: BEMMERL. Thomas. MPICH for SCI-connected clusters. In: SCI-EUROPE. 1999, Toulouse. France. Proccedings ... [S.I.: s.n.]. 1999. p.3-11.
WORRINGEN, Joachirn. SCI-MPICH: the second generation. In: SCI EUROPE, 3., 2000, Munich, Germany. Proceedings ... [S.I.: s.n.], 2000. p. 10-20. Organizado como conference stream do Euro-Par'2000.
ZORAJA, Ivan; HELLWAGNER, Herrmann; SUNDERAM. Vaidy. SCIPVM: parallel distributed cornputing on SCI workstation clusters. Concurrency: Practice and Experience. v.II. n.13, p. 121-138, Mar. 1999.
Publicado
10/09/2001
Como Citar
OLIVEIRA, Fábio A. D.; ÁVILA, Rafael B.; BARRETO, Marcos E.; NAVAUX, Philippe O. A..
Low-latency and zero-copy message passing protocols for SCI-based clusters. In: INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE AND HIGH PERFORMANCE COMPUTING (SBAC-PAD), 13. , 2001, Pirenópolis.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2001
.
p. 148-155.
DOI: https://doi.org/10.5753/sbac-pad.2001.22203.
