Energy and Performance Prediction of CUDA Applications using Dynamic Regression Models

被引:2
|
作者
Benedict, Shajulin [1 ]
Rejitha, R. S. [1 ]
Alex, Suja A. [1 ]
机构
[1] Anna Univ, SXCCE, HPCCLoud Res Lab, Madras 600025, Tamil Nadu, India
关键词
Applications; CUDA; Energy; Performance Tuning; Performance Analysis; Tools; TIMES;
D O I
10.1145/2856636.2856643
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Many emerging supercomputers and future exa-scale computing machines require accelerator-based GPU computing architectures for boosting their computing performances. CUDA is one of the widely applied GPGPU parallel computing platform for those architectures owing to its better performance for certain scientific applications. However, the emerging rise in the development of CUDA applications from various scientific domains, such as, bio-informatics, HEP, and so forth, has urged the need for tools that identify optimal application parameters and the other GPGPU architecture metrics, including work group size, work item, memory utilization, and so forth. In fact, the tuning process might end up with several executions of various possible code variants. This paper proposed Dynamic Regression models, namely, Dynamic Random Forests (DynRFM), Dynamic Support Vector Machines (DynSVM), and Dynamic Linear Regression Models (Dyn LRM) for the energy/performance prediction of the code variants of CUDA applications. The prediction was based on the application parameters and the performance metrics of applications, such as, number of instructions, memory issues, and so forth. In order to obtain energy/performance measurements for CUDA applications, EACudaLib (a monitoring library implemented in EnergyAnalyzer tool) was developed. In addition, the proposed Dynamic Regression models were compared to the classical regression models, such as, RFM, SVM, and LRM. The validation results of the proposed dynamic regression models, when tested with the different problem sizes of Nbody and Particle CUDA simulations, manifested the energy/performance prediction improvement of over 50.26 to 61.23 percentages.
引用
收藏
页码:37 / 47
页数:11
相关论文
共 50 条
  • [1] Energy prediction of CUDA application instances using dynamic regression models
    Rejitha, R. S.
    Benedict, Shajulin
    Alex, Suja A.
    Infanto, Shany
    COMPUTING, 2017, 99 (08) : 765 - 790
  • [2] Energy prediction of CUDA application instances using dynamic regression models
    R. S. Rejitha
    Shajulin Benedict
    Suja A. Alex
    Shany Infanto
    Computing, 2017, 99 : 765 - 790
  • [3] Dynamic Regression Models for Prediction of Construction Costs
    Hwang, Seokyon
    JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT-ASCE, 2009, 135 (05): : 360 - 367
  • [4] Performance Measurement of Applications with GPU Acceleration using CUDA
    Mayanglambam, Shangkar
    Malony, Allen D.
    Sottile, Matthew J.
    PARALLEL COMPUTING: FROM MULTICORES AND GPU'S TO PETASCALE, 2010, 19 : 341 - 348
  • [5] Labour Productivity Prediction Estimate in Manufacturing Environment using Regression and Dynamic Models
    Abdenov, Amirza
    Abdenova, Gaukhar
    Kenzhebayeva, Zhanat
    Mukhanova, Ayagoz
    MATERIALS TODAY-PROCEEDINGS, 2019, 16 : 254 - 261
  • [6] Development of performance prediction models in flexible pavement using regression analysis method
    Sung-Hee Kim
    Nakseok Kim
    KSCE Journal of Civil Engineering, 2006, 10 (2) : 91 - 96
  • [7] ANALYSIS AND PREDICTION OF TRANSMISSION UNIT PERFORMANCE USING ADVANCED REGRESSION-MODELS
    SCHNEIDER, AW
    LANDGREN, GL
    BHAVARAJU, MP
    IEEE TRANSACTIONS ON POWER APPARATUS AND SYSTEMS, 1985, 104 (05): : 1084 - 1094
  • [8] Performance evaluation of unified memory and dynamic parallelism for selected parallel CUDA applications
    Jarzabek, Lukasz
    Czarnul, Pawel
    JOURNAL OF SUPERCOMPUTING, 2017, 73 (12): : 5378 - 5401
  • [9] Performance Prediction Framework for CUDA Programs
    Qu H.-C.
    Yu S.-M.
    Liu W.-J.
    Wang X.-Y.
    Tien Tzu Hsueh Pao/Acta Electronica Sinica, 2020, 48 (04): : 654 - 661
  • [10] Performance evaluation of unified memory and dynamic parallelism for selected parallel CUDA applications
    Łukasz Jarząbek
    Paweł Czarnul
    The Journal of Supercomputing, 2017, 73 : 5378 - 5401