Abstract
With the increasing diversity of application needs and computing units, the server with heterogeneous processors is more and more widespread. However, conventional SMP/ccNUMA server architecture introduces communication bottleneck between heterogeneous processors and only uses heterogeneous processors as coprocessors, which limits the efficiency and flexibility of using heterogeneous processors. To solve this problem, this paper proposes an intra-server interconnect fabric that supports both intra-server peer-to-peer interconnection and I/O resource sharing among heterogeneous processors. By connecting processors and I/O devices with the proposed fabric, heterogeneous processors can perform direct communication with each other and run in stand-alone mode with shared intra-server resources. We design the proposed fabric by extending the de-facto system I/O bus protocol PCIe (Peripheral Computer Interconnect Express) and implement it with a single chip cZodiac. By making full use of PCIe’s original advantages, the interconnection and the I/O sharing mechanism are light weight and efficient. Evaluations that have been carried out on both the FPGA (Field Programmable Gate Array) prototype and the cycle-accurate simulator demonstrate that our design is feasible and scalable. In addition, our design is suitable for not only the heterogeneous server but also the high density server.
Similar content being viewed by others
Explore related subjects
Discover the latest articles and news from researchers in related subjects, suggested using machine learning.References
Barker K J, Davis K, Hoisie A et al. Entering the petaflop era: The architecture and performance of Roadrunner. In Proc. ACM/IEEE Conf. Supercomputing, Nov. 2008, Article No. 1.
Sun N H, Xing J, Huo Z G et al. Dawning Nebulae: A petaFLOPS supercomputer with a heterogeneous structure. Journal of Computer Science and Technology, 2011, 26(3): 352–362.
Reddi V J, Lee B C, Chilimbi T, Vaid K. Web search using mobile cores: Quantifying and mitigating the price of efficiency. In Proc. the 37th Annual Int. Symp. Computer Architecture, June 2010, pp.314–325.
Guevara M, Lubin B, Lee B C. Navigating heterogeneous processors with market mechanisms. In Proc. the 19th IEEE Int. Symp. High Performance Computer Architecture, Feb. 2013, pp.95–106.
Zapater M, Ayala J L, Moya J M. Leveraging heterogeneity for energy minimization in data centers. In Proc. the 12th IEEE/ACM Int. Symp. Cluster, Cloud and Grid Computing, May 2012, pp.752–757.
Suneja S, Baron E, Lara E D et al. Accelerating the cloud with heterogeneous computing. In Proc. the 3rd USENIX Conf. Hot Topics in Cloud Computing, June 2011, p.23.
Peh L S, Dally W J. A delay model and speculative architecture for pipelined routers. In Proc. the 7th Int. Symp. High Performance Computer Architecture, Jan. 2001, pp.255–266.
Ohno Y, Nishibori E, Narumi T et al. A 281Tflops calculation for X-ray protein structure analysis with special-purpose computers MDGRAPE-3. In Proc. ACM/IEEE Conference on Supercomputing, Nov. 2007, pp.1–10.
Wong D, Annavaram M. KnightShift: Scaling the energy proportionality wall through server-level heterogeneity. In Proc. the 45th IEEE/ACM Int. Symp. Microarchitecture, Dec. 2012, pp.119–130.
Krishnan V. Evaluation of an integrated PCI express IO expansion and clustering fabric. In Proc. the 16th IEEE Symp. High Performance Interconnects, Aug. 2008, pp.93–100.
Krishnan V. Towards an integrated IO and clustering solution using PCI express. In Proc. IEEE International Conference on Cluster Computing, Sept. 2007, pp.259–266.
Aswadhati A. Scaling data center services with PCI express. In Proc. Linley Tech. Data Center Conference, Feb. 2012.
Suzuki J, Hidaka Y, Higuchi J et al. Multi-root share of single-root I/O virtualization (SR-IOV) compliant PCI Express device. In Proc. the 18th IEEE Symp. High Performance Interconnects, Aug. 2010. pp.25–31.
Author information
Authors and Affiliations
Corresponding author
Additional information
This work was supported by the National Natural Science Foundation of China under Grant No. 61100014.
Electronic supplementary material
Below is the link to the electronic supplementary material.
ESM 1
(PDF 84 kb)
Rights and permissions
About this article
Cite this article
Cao, Z., Liu, XL., Li, Q. et al. An Intra-Server Interconnect Fabric for Heterogeneous Computing. J. Comput. Sci. Technol. 29, 976–988 (2014). https://doi.org/10.1007/s11390-014-1483-0
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11390-014-1483-0