Abstract
The TI Keystone II architecture provides a unique combination of ARM Cortex-A15 processors with high performance TI C66x floating-point DSPs on a single low-power System-on-chip (SoC). Commercially available systems such as the HP Proliant m800 and nCore BrownDwarf are based on this ARM-DSP SoC. The Keystone II architecture promises to deliver high GFLOPS/Watt and is of increasing interest as it provides an alternate building block for future exascale systems. However, the success of this architecture is intimately related to the ease of migrating existing HPC applications for maximum performance. Effective use of all ARM and DSP cores and DMA co-processors is crucial for maximizing performance/watt. This paper explores issues and challenges encountered while migrating the matrix multiplication (GEMM) kernel, originally written only for the C6678 DSP to the ARM-DSP SoC using an early prototype of the OpenMP 4.0 accelerator model. Single precision (SGEMM) matrix multiplication performance of 110.11 GFLOPS and and double precision (DGEMM) performance of 29.15 GFLOPS was achieved on the TI Keystone II Evaluation Module Revision 3.0 (EVM). Trade-offs and factors affecting performance are discussed.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Ali, M., Stotzer, E., Igual, F.D., van de Geijn, R.A.: Level-3 BLAS on the TI C6678 multi-core DSP. In: IEEE 24th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD), pp. 179–186. IEEE (2012)
Igual, F.D., Ali, M., Friedmann, A., Stotzer, E., Wentz, T., van de Geijn, R.A.: Unleashing the high-performance and low-power of multi-core DSPs for general-purpose HPC. In: Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis, p. 26. IEEE Computer Society Press (2012)
HP: HP moonshot system (2014), http://h17007.www1.hp.com/us/en/enterprise/servers/products/moonshot/index.aspx
nCore HPC: ncore browndwarf y-class supercomputer (2014), http://ncorehpc.com/browndwarf/
Stotzer, E., Jayaraj, A., Ali, M., Friedmann, A., Mitra, G., Rendell, A.P., Lintault, I.: OpenMP on the Low-Power TI Keystone II ARM/DSP System-on-Chip. In: Rendell, A.P., Chapman, B.M., Müller, M.S. (eds.) IWOMP 2013. LNCS, vol. 8122, pp. 114–127. Springer, Heidelberg (2013)
OpenMP ARB: OpenMP Application Program Interface, v.4.0 (2013), http://www.openmp.org/mp-documents/OpenMP4.0.0.pdf
Texas Instruments Literature: SPRS866: 66AK2H12/06 Multicore DSP+ARM Keystone II System-on-Chip (SoC)
Rajovic, N., Rico, A., Puzovic, N., Adeniyi-Jones, C., Ramirez, A.: Tibidabo: making the case for an ARM-based HPC system (2013)
Mitra, G., Johnston, B., Rendell, A.P., McCreath, E., Zhou, J.: Use of SIMD vector operations to accelerate application code performance on low-powered ARM and Intel platforms. In: Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW). IEEE (2013)
Khronos: OpenCL: The open standard for parallel programming of heterogeneous systems (2011), http://www.khronos.org/opencl
Reyes, R., Lopez, I., Fumero, J.J., de Sande, F.: Directive-based programming for gpus: A comparative study. In: IEEE 14th International Conference on High Performance Computing and Communication & 2012 IEEE 9th International Conference on Embedded Software and Systems (HPCC-ICESS), pp. 410–417. IEEE (2012)
Wolfe, M.: Implementing the PGI accelerator model. In: Proceedings of the 3rd Workshop on General-Purpose Computation on Graphics Processing Units, pp. 43–50. ACM (2010)
Han, T.D., Abdelrahman, T.S.: Hi CUDA: A high-level directive-based language for GPU programming. In: Proceedings of 2nd Workshop on General Purpose Processing on Graphics Processing Units, pp. 52–61. ACM (2009)
Ahmad, A., Ali, M., South, F., Monroy, G.L., Adie, S.G., Shemonski, N., Carney, P.S., Boppart, S.A.: Interferometric synthetic aperture microscopy implementation on a floating point multi-core digital signal processer. In: SPIE BiOS, International Society for Optics and Photonics, pp. 857134–857134 (2013)
Note, F.W., Van Zee, F.G., Smith, T., Igual, F.D., Smelyanskiy, M., Zhang, X., Kistler, M., Austel, V., Gunnels, J., Low, T.M., et al.: Implementing level-3 blas with blis: Early experience (2013)
NVIDIA: Unified Memory in CUDA 6 (2014), http://devblogs.nvidia.com/parallelforall/unified-memory-in-cuda-6/
NVIDIA: NVIDIA Tegra K1 Processor (2014), http://www.nvidia.com/object/tegra-k1-processor.html
Liao, C., Yan, Y., de Supinski, B.R., Quinlan, D.J., Chapman, B.: Early Experiences With The OpenMP Accelerator Model. In: Rendell, A.P., Chapman, B.M., Müller, M.S. (eds.) IWOMP 2013. LNCS, vol. 8122, pp. 84–98. Springer, Heidelberg (2013)
Schmidl, D., Cramer, T., Wienke, S., Terboven, C., Müller, M.S.: Assessing the performance of OpenMP programs on the Intel Xeon Phi. In: Wolf, F., Mohr, B., an Mey, D. (eds.) Euro-Par 2013. LNCS, vol. 8097, pp. 547–558. Springer, Heidelberg (2013)
Barker, J., Bowden, J.: Manycore Parallelism through OpenMP. In: Rendell, A.P., Chapman, B.M., Müller, M.S. (eds.) IWOMP 2013. LNCS, vol. 8122, pp. 45–57. Springer, Heidelberg (2013)
Cramer, T., Schmidl, D., Klemm, M., an Mey, D.: OpenMP Programming on Intel Xeon Phi Coprocessors: An Early Performance Comparison, pp. 38–44 (2012)
Leang, S.S., Rendell, A.P., Gordon, M.S.: Quantum chemical calculations using accelerators: Migrating matrix operations to the nvidia kepler gpu and the intel xeon phi. Journal of Chemical Theory and Computation 10(3), 908–912 (2014)
Newburn, C., Dmitriev, S., Narayanaswamy, R., Wiegert, J., Murty, R., Chinchilla, F., Deodhar, R., McGuire, R.: Offload Compiler Runtime for the Intel Xeon Phi Coprocessor. In: 2013 IEEE 27th International on Parallel and Distributed Processing Symposium Workshops PhD Forum (IPDPSW), pp. 1213–1225 (May 2013)
Li, B., Chang, H.C., Leon Song, S., Su, C.Y., Meyer, T., Mooring, J., Cameron, K.W.: The Power-Performance Tradeoffs of the Intel Xeon Phi on HPC Applications. In: Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW). IEEE (2014)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Mitra, G., Stotzer, E., Jayaraj, A., Rendell, A.P. (2014). Implementation and Optimization of the OpenMP Accelerator Model for the TI Keystone II Architecture. In: DeRose, L., de Supinski, B.R., Olivier, S.L., Chapman, B.M., Müller, M.S. (eds) Using and Improving OpenMP for Devices, Tasks, and More. IWOMP 2014. Lecture Notes in Computer Science, vol 8766. Springer, Cham. https://doi.org/10.1007/978-3-319-11454-5_15
Download citation
DOI: https://doi.org/10.1007/978-3-319-11454-5_15
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-11453-8
Online ISBN: 978-3-319-11454-5
eBook Packages: Computer ScienceComputer Science (R0)