DROPS

Document

DOI: 10.4230/LIPIcs.ECRTS.2018.19

Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms

Authors: Waqar Ali and Heechul Yun

Published in: LIPIcs, Volume 106, 30th Euromicro Conference on Real-Time Systems (ECRTS 2018)

Abstract

Integrated CPU-GPU architecture provides excellent acceleration capabilities for data parallel applications on embedded platforms while meeting the size, weight and power (SWaP) requirements. However, sharing of main memory between CPU applications and GPU kernels can severely affect the execution of GPU kernels and diminish the performance gain provided by GPU. For example, in the NVIDIA Jetson TX2 platform, an integrated CPU-GPU architecture, we observed that, in the worst case, the GPU kernels can suffer as much as 3X slowdown in the presence of co-running memory intensive CPU applications. In this paper, we propose a software mechanism, which we call BWLOCK++, to protect the performance of GPU kernels from co-scheduled memory intensive CPU applications.

Cite as

Waqar Ali and Heechul Yun. Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms. In 30th Euromicro Conference on Real-Time Systems (ECRTS 2018). Leibniz International Proceedings in Informatics (LIPIcs), Volume 106, pp. 19:1-19:22, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2018)

Copy BibTex To Clipboard

@InProceedings{ali_et_al:LIPIcs.ECRTS.2018.19,
  author =	{Ali, Waqar and Yun, Heechul},
  title =	{{Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms}},
  booktitle =	{30th Euromicro Conference on Real-Time Systems (ECRTS 2018)},
  pages =	{19:1--19:22},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-075-0},
  ISSN =	{1868-8969},
  year =	{2018},
  volume =	{106},
  editor =	{Altmeyer, Sebastian},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECRTS.2018.19},
  URN =		{urn:nbn:de:0030-drops-89833},
  doi =		{10.4230/LIPIcs.ECRTS.2018.19},
  annote =	{Keywords: GPU, memory bandwidth, resource contention, CPU throttling, fair scheduler}
}

Document

DOI: 10.4230/DARTS.4.2.3

Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms (Artifact)

Authors: Waqar Ali and Heechul Yun

Published in: DARTS, Volume 4, Issue 2, Special Issue of the 30th Euromicro Conference on Real-Time Systems (ECRTS 2018)

Abstract

This artifact is based on BWLOCK++, a software framework to protect the performance of GPU kernels from co-scheduled memory intensive CPU applications in platforms containing integrated GPUs. The artifact is designed to support the claims of the companion paper and contains instructions on how to build and execute BWLOCK++ on a target hardware platform.

Cite as

Waqar Ali and Heechul Yun. Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms (Artifact). In Special Issue of the 30th Euromicro Conference on Real-Time Systems (ECRTS 2018). Dagstuhl Artifacts Series (DARTS), Volume 4, Issue 2, pp. 3:1-3:2, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2018)

Copy BibTex To Clipboard

@Article{ali_et_al:DARTS.4.2.3,
  author =	{Ali, Waqar and Yun, Heechul},
  title =	{{Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms (Artifact)}},
  pages =	{3:1--3:2},
  journal =	{Dagstuhl Artifacts Series},
  ISSN =	{2509-8195},
  year =	{2018},
  volume =	{4},
  number =	{2},
  editor =	{Ali, Waqar and Yun, Heechul},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DARTS.4.2.3},
  URN =		{urn:nbn:de:0030-drops-89719},
  doi =		{10.4230/DARTS.4.2.3},
  annote =	{Keywords: GPU, memory bandwidth, resource contention, CPU throttling, fair scheduler}
}

Search Results

Documents authored by Ali, Waqar

Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms

Abstract

Cite as

Protecting Real-Time GPU Kernels on Integrated CPU-GPU SoC Platforms (Artifact)

Abstract

Cite as

Thanks for your feedback!

Could not send message