Search Results

Documents authored by Kosmidis, Leonidas

Document
DOI: 10.4230/LIPIcs.ECRTS.2019.23
Generating and Exploiting Deep Learning Variants to Increase Heterogeneous Resource Utilization in the NVIDIA Xavier

Authors: Roger Pujol, Hamid Tabani, Leonidas Kosmidis, Enrico Mezzetti, Jaume Abella, and Francisco J. Cazorla

Published in: LIPIcs, Volume 133, 31st Euromicro Conference on Real-Time Systems (ECRTS 2019)

Abstract
Deep learning-based solutions and, in particular, deep neural networks (DNNs) are at the heart of several functionalities in critical-real time embedded systems (CRTES) from vision-based perception (object detection and tracking) systems to trajectory planning. As a result, several DNN instances simultaneously run at any time on the same computing platform. However, while modern GPUs offer a variety of computing elements (e.g. CPUs, GPUs, and specific accelerators) in which those DNN tasks can be executed depending on their computational requirements and temporal constraints, current DNNs are mainly programmed to exploit one of them, namely, regular cores in the GPU. This creates resource imbalance and under-utilization of GPU resources when executing several DNN instances, causing an increase in DNN tasks' execution time requirements. In this paper, (a) we develop different variants (implementations) of well-known DNN libraries used in the Apollo Autonomous Driving (AD) software for each of the computing elements of the latest NVIDIA Xavier SoC. Each variant can be configured to balance resource requirements and performance: the regular CPU core implementation that can run on 2, 4, and 6 cores; the GPU regular and Tensor core variants that can run in 4 or 8 GPU’s Streaming Multiprocessors (SM); and 1 or 2 NVIDIA’s Deep Learning Accelerators (NVDLA); (b) we show that each particular variant/configuration offers a different resource utilization/performance point; finally, (c) we show how those heterogeneous computing elements can be exploited by a static scheduler to sustain the execution of multiple and diverse DNN variants on the same platform.
Cite as

Roger Pujol, Hamid Tabani, Leonidas Kosmidis, Enrico Mezzetti, Jaume Abella, and Francisco J. Cazorla. Generating and Exploiting Deep Learning Variants to Increase Heterogeneous Resource Utilization in the NVIDIA Xavier. In 31st Euromicro Conference on Real-Time Systems (ECRTS 2019). Leibniz International Proceedings in Informatics (LIPIcs), Volume 133, pp. 23:1-23:23, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2019)

Copy BibTex To Clipboard

@InProceedings{pujol_et_al:LIPIcs.ECRTS.2019.23,
  author =	{Pujol, Roger and Tabani, Hamid and Kosmidis, Leonidas and Mezzetti, Enrico and Abella, Jaume and Cazorla, Francisco J.},
  title =	{{Generating and Exploiting Deep Learning Variants to Increase Heterogeneous Resource Utilization in the NVIDIA Xavier}},
  booktitle =	{31st Euromicro Conference on Real-Time Systems (ECRTS 2019)},
  pages =	{23:1--23:23},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-110-8},
  ISSN =	{1868-8969},
  year =	{2019},
  volume =	{133},
  editor =	{Quinton, Sophie},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ECRTS.2019.23},
  URN =		{urn:nbn:de:0030-drops-107608},
  doi =		{10.4230/LIPIcs.ECRTS.2019.23},
  annote =	{Keywords: Deep Neural Network (DNN), GPU, Heterogenous Resources}
}
Document
DOI: 10.4230/OASIcs.WCET.2016.9
Measurement-Based Timing Analysis of the AURIX Caches

Authors: Leonidas Kosmidis, Davide Compagnin, David Morales, Enrico Mezzetti, Eduardo Quinones, Jaume Abella, Tullio Vardanega, and Francisco J. Cazorla

Published in: OASIcs, Volume 55, 16th International Workshop on Worst-Case Execution Time Analysis (WCET 2016)

Abstract
Cache memories are one of the hardware resources with higher potential to reduce worst-case execution time (WCET) costs for software programs with tight real-time constraints. Yet, the complexity of cache analysis has caused a large fraction of real-time systems industry to avoid using them, especially in the automotive sector. For measurement-based timing analysis (MBTA) - the dominant technique in domains such as automotive - cache challenges the definition of test scenarios stressful enough to produce (cache) layouts that causing high contention. In this paper, we present our experience in enabling the use of caches for a real automotive application running on an AURIX multiprocessor, using software randomization and measurement-based probabilistic timing analysis (MBPTA). Our results show that software randomization successfully exposes - in the experiments performed for timing analysis - cache related variability, in a manner that can be effectively captured by MBPTA.
Cite as

Leonidas Kosmidis, Davide Compagnin, David Morales, Enrico Mezzetti, Eduardo Quinones, Jaume Abella, Tullio Vardanega, and Francisco J. Cazorla. Measurement-Based Timing Analysis of the AURIX Caches. In 16th International Workshop on Worst-Case Execution Time Analysis (WCET 2016). Open Access Series in Informatics (OASIcs), Volume 55, pp. 9:1-9:11, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2016)

Copy BibTex To Clipboard

@InProceedings{kosmidis_et_al:OASIcs.WCET.2016.9,
  author =	{Kosmidis, Leonidas and Compagnin, Davide and Morales, David and Mezzetti, Enrico and Quinones, Eduardo and Abella, Jaume and Vardanega, Tullio and Cazorla, Francisco J.},
  title =	{{Measurement-Based Timing Analysis of the AURIX Caches}},
  booktitle =	{16th International Workshop on Worst-Case Execution Time Analysis (WCET 2016)},
  pages =	{9:1--9:11},
  series =	{Open Access Series in Informatics (OASIcs)},
  ISBN =	{978-3-95977-025-5},
  ISSN =	{2190-6807},
  year =	{2016},
  volume =	{55},
  editor =	{Schoeberl, Martin},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/OASIcs.WCET.2016.9},
  URN =		{urn:nbn:de:0030-drops-69028},
  doi =		{10.4230/OASIcs.WCET.2016.9},
  annote =	{Keywords: WCET, caches, AURIX, Automotive}
}
Document
DOI: 10.4230/OASIcs.WCET.2013.97
Applying Measurement-Based Probabilistic Timing Analysis to Buffer Resources

Authors: Leonidas Kosmidis, Tullio Vardanega, Jaume Abella, Eduardo Quiñones, and Francisco J. Cazorla

Published in: OASIcs, Volume 30, 13th International Workshop on Worst-Case Execution Time Analysis (2013)

Abstract
The use of complex hardware makes it difficult for current timing analysis techniques to compute trustworthy and tight worst-case execution time (WCET) bounds. Those techniques require detailed knowledge of the internal operation and state of the platform, at both the software and hardware level. Obtaining that information for modern hardware platforms is increasingly difficult. Measurement-Based Probabilistic Timing Analysis (MBPTA) reduces the cost of acquiring the knowledge needed for computing trustworthy and tight WCET bounds. MBPTA based on Extreme Value Theory requires the execution time of processor instructions to be independent and identically distributed (i.i.d.), which can be achieved with some hardware support. Previous proposals show how those properties can be achieved for caches. This paper considers, for the first time, the implications on MBPTA of using buffer resources. Buffers in general, and first-come first-served (FCFS) buffers in particular, are of paramount importance as the complexity of hardware increases, since they allow managing contention in those resources where multiple requests may be pending. We show how buffers can be used in the context of MBPTA and provide illustrative examples.
Cite as

Leonidas Kosmidis, Tullio Vardanega, Jaume Abella, Eduardo Quiñones, and Francisco J. Cazorla. Applying Measurement-Based Probabilistic Timing Analysis to Buffer Resources. In 13th International Workshop on Worst-Case Execution Time Analysis. Open Access Series in Informatics (OASIcs), Volume 30, pp. 97-108, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2013)

Copy BibTex To Clipboard

@InProceedings{kosmidis_et_al:OASIcs.WCET.2013.97,
  author =	{Kosmidis, Leonidas and Vardanega, Tullio and Abella, Jaume and Qui\~{n}ones, Eduardo and Cazorla, Francisco J.},
  title =	{{Applying Measurement-Based Probabilistic Timing Analysis to Buffer Resources}},
  booktitle =	{13th International Workshop on Worst-Case Execution Time Analysis},
  pages =	{97--108},
  series =	{Open Access Series in Informatics (OASIcs)},
  ISBN =	{978-3-939897-54-5},
  ISSN =	{2190-6807},
  year =	{2013},
  volume =	{30},
  editor =	{Maiza, Claire},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/OASIcs.WCET.2013.97},
  URN =		{urn:nbn:de:0030-drops-41269},
  doi =		{10.4230/OASIcs.WCET.2013.97},
  annote =	{Keywords: WCET, Buffer, Probabilistic Timing Analysis}
}
Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail
© 2023-2024 Schloss Dagstuhl – LZI GmbH About DROPS Imprint Privacy Contact