Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

GMAI

GMAI Critical real-time systems require strict resource provisioning in terms of memory and timing. The constant need for higher performance in these systems has led industry to recently include GPUs. However, GPU software ecosystems are by their nature closed source, forcing system engineers to consider them as black boxes, complicating resource provisioning. In this work, we reverse engineer the internal operations of the GPU system software to increase the understanding of their observed behaviour and how resources are internally managed. We present our methodology that is incorporated in GMAI (GPU Memory Allocation Inspector), a tool that allows system engineers to accurately determine the exact amount of resources required by their critical systems, avoiding underprovisioning. We first apply our methodology on a wide range of GPU hardware from different vendors showing its generality in obtaining the properties of the GPU memory allocators. Next, we demonstrate the benefits of such knowledge in resource provisioning of two case studies from the automotive domain, where the actual memory consumption is up to 5.6× more than the memory requested by the application. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ACM Transactions on Embedded Computing Systems (TECS) Association for Computing Machinery

GMAI

Download PDF
23 pages

Loading next page...
 
/lp/association-for-computing-machinery/gmai-KLg47Y8E1T
Publisher
Association for Computing Machinery
Copyright
Copyright © 2020 ACM
ISSN
1539-9087
eISSN
1558-3465
DOI
10.1145/3391896
Publisher site
See Article on Publisher Site

Abstract

Critical real-time systems require strict resource provisioning in terms of memory and timing. The constant need for higher performance in these systems has led industry to recently include GPUs. However, GPU software ecosystems are by their nature closed source, forcing system engineers to consider them as black boxes, complicating resource provisioning. In this work, we reverse engineer the internal operations of the GPU system software to increase the understanding of their observed behaviour and how resources are internally managed. We present our methodology that is incorporated in GMAI (GPU Memory Allocation Inspector), a tool that allows system engineers to accurately determine the exact amount of resources required by their critical systems, avoiding underprovisioning. We first apply our methodology on a wide range of GPU hardware from different vendors showing its generality in obtaining the properties of the GPU memory allocators. Next, we demonstrate the benefits of such knowledge in resource provisioning of two case studies from the automotive domain, where the actual memory consumption is up to 5.6× more than the memory requested by the application.

Journal

ACM Transactions on Embedded Computing Systems (TECS) – Association for Computing Machinery

Published: Sep 26, 2020

Keywords: GPUs

References

  • GPU scheduling on the NVIDIA TX2: Hidden details revealed
    Amert, T.; Otterness, N.; Yang, M.; Anderson, J. H.; Smith, F. Donelson
  • Avionics Application Software Standard Interface: ARINC Specification 653P1-3
    ARINC.,
  • AUTOSAR
    AUTOSAR.,
  • Hoard: A scalable memory allocator for multithreaded applications
    Berger, E. D.; McKinley, K. S.; Blumofe, R. D.; Wilson, P. R.
  • GMAI: GPU Memory Allocation Inspector
    Calderón, A. J.; Kosmidis, L.; Nicolás, C. F.; Cazorla, F. J.; Onaindia, P.
  • Who allocated my memory? Detecting custom memory allocators in C binaries
    Chen, X.; Slowinska, A.; Bos, H.
  • Error resilient GPU accelerated image processing for space applications
    Davidson, R. L.; Bridges, C. P.
  • The GNU Allocator
    Foundation, Free Software
  • Integrity RTOS
    Software, Green Hills
  • A tunable hybrid memory allocator
    Hasan, Y.; Chang, J. M.
  • XMalloc: A scalable lock-free dynamic memory allocator for many-core machines
    Huang, X.; Rodrigues, C. I.; Jones, S.; Buck, I.; Hwu, W.
  • Hwu
    Huang, X.; Rodrigues, C. I.; Jones, S.; Buck, I.; W.-m,
  • Getting the Most from OpenCL 1
    Corporation, Intel
  • Industrial experiences with resource management under software randomization in ARINC653 avionics environments
    Kosmidis, L.; Maxim, C.; Jegu, V.; Vatrinet, F.; Cazorla, F. J.
  • Dissecting GPU memory hierarchy through microbenchmarking
    Mei, X.; Chu, X.
  • Self Driving Cars
    Corporation, NVIDIA
  • Combination of the symmetrical local threshold and the sobel edge detector for lane feature extraction
    Ozgunalp, U.
  • Proposed Memory Allocation Algorithmfor NUMA-based Soft Real-Time Operating System
    Shah, V.; Shah, A.
  • Howard: A dynamic excavator for reverse engineering data structures
    Slowinska, A.; Stancescu, T.; Bos, H.
  • ScatterAlloc: Massively parallel dynamic memory allocation for the GPU
    Steinberger, M.; Kenzel, M.; Kainz, B.; Schmalstieg, D.
  • An open benchmark implementation for multi-CPU multi-GPU pedestrian detection in automotive systems
    Trompouki, M. M.; Kosmidis, L.; Navarro, N.
  • Traffic sign recognition in autonomous vehicles using edge detection
    Vishwanathan, H.; Peters, D. L.; Zhang, J. Z.
  • Fast dynamic memory allocator for massively parallel architectures
    Widmer, S.; Wodniok, D.; Weber, N.; Goesele, M.
  • Dynamic Storage Allocation: A Survey and Critical Review
    Wilson, P. R.; Johnstone, M. S.; Neely, M.; Boles, D.
  • Demystifying GPU microarchitecture through microbenchmarking
    Wong, H.; Papadopoulou, M.; Sadooghi-Alvandi, M.; Moshovos, A.
  • Avoiding pitfalls when using NVIDIA GPUs for real-time tasks in autonomous systems
    Yang, M.; Otterness, N.; Amert, T.; Bakita, J.; Anderson, J. H.; Smith, F. D.
  • Accelerated fog removal from real images for car detection
    Younis, R.; Bastaki, N.
  • GPU-based iterative medical CT image reconstructions
    Yu, X.; Wang, H.; Feng, W.; Gong, H.; Cao, G.