<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en-gb">
	<link rel="self" type="application/atom+xml" href="http://localhost/app.php/feed/topic/124" />

	<title>Tools and Benchmarks for Real-Time Systems</title>
	<subtitle>ECRTS Community Forum</subtitle>
	<link href="http://localhost/index.php" />
	<updated>2019-03-26T08:23:14+01:00</updated>

	<author><name><![CDATA[Tools and Benchmarks for Real-Time Systems]]></name></author>
	<id>http://localhost/app.php/feed/topic/124</id>

		<entry>
		<author><name><![CDATA[arne.hamann]]></name></author>
		<updated>2019-03-26T08:23:14+01:00</updated>

		<published>2019-03-26T08:23:14+01:00</published>
		<id>http://localhost/viewtopic.php?t=124&amp;p=239#p239</id>
		<link href="http://localhost/viewtopic.php?t=124&amp;p=239#p239"/>
		<title type="html"><![CDATA[Re: Description of the WATERS Industrial Challenge 2019]]></title>

		
		<content type="html" xml:base="http://localhost/viewtopic.php?t=124&amp;p=239#p239"><![CDATA[
<blockquote class="uncited"><div><blockquote class="uncited"><div>I am glad to announce the description of the WATERS Industrial Challenge 2019 that is proposed by Bosch in cooperation with University of Modena.<br><br>Please note that the Amalthea performance file that is necessary to respond to the challenge will be delivered in 2-3 weeks.</div></blockquote>Any update on this?</div></blockquote>The model is now available!<p>Statistics: Posted by <a href="http://localhost/memberlist.php?mode=viewprofile&amp;u=708">arne.hamann</a> — Tue Mar 26, 2019</p><hr />
]]></content>
	</entry>
		<entry>
		<author><name><![CDATA[zero212]]></name></author>
		<updated>2019-03-23T14:23:30+01:00</updated>

		<published>2019-03-23T14:23:30+01:00</published>
		<id>http://localhost/viewtopic.php?t=124&amp;p=237#p237</id>
		<link href="http://localhost/viewtopic.php?t=124&amp;p=237#p237"/>
		<title type="html"><![CDATA[Re: Description of the WATERS Industrial Challenge 2019]]></title>

		
		<content type="html" xml:base="http://localhost/viewtopic.php?t=124&amp;p=237#p237"><![CDATA[
<blockquote class="uncited"><div>I am glad to announce the description of the WATERS Industrial Challenge 2019 that is proposed by Bosch in cooperation with University of Modena.<br><br>Please note that the Amalthea performance file that is necessary to respond to the challenge will be delivered in 2-3 weeks.</div></blockquote>Any update on this?<p>Statistics: Posted by <a href="http://localhost/memberlist.php?mode=viewprofile&amp;u=25212">zero212</a> — Sat Mar 23, 2019</p><hr />
]]></content>
	</entry>
		<entry>
		<author><name><![CDATA[arne.hamann]]></name></author>
		<updated>2019-03-06T18:22:29+01:00</updated>

		<published>2019-03-06T18:22:29+01:00</published>
		<id>http://localhost/viewtopic.php?t=124&amp;p=236#p236</id>
		<link href="http://localhost/viewtopic.php?t=124&amp;p=236#p236"/>
		<title type="html"><![CDATA[Re: Description of the WATERS Industrial Challenge 2019]]></title>

		
		<content type="html" xml:base="http://localhost/viewtopic.php?t=124&amp;p=236#p236"><![CDATA[
I just uploaded an updated version of the description. We forgot to mention that the response time analysis should be memory contention aware. There is an example of a possible model already in the forum.<p>Statistics: Posted by <a href="http://localhost/memberlist.php?mode=viewprofile&amp;u=708">arne.hamann</a> — Wed Mar 06, 2019</p><hr />
]]></content>
	</entry>
		<entry>
		<author><name><![CDATA[Nacho_S]]></name></author>
		<updated>2019-03-05T12:36:32+01:00</updated>

		<published>2019-03-05T12:36:32+01:00</published>
		<id>http://localhost/viewtopic.php?t=124&amp;p=234#p234</id>
		<link href="http://localhost/viewtopic.php?t=124&amp;p=234#p234"/>
		<title type="html"><![CDATA[Re: Description of the WATERS Industrial Challenge 2019]]></title>

		
		<content type="html" xml:base="http://localhost/viewtopic.php?t=124&amp;p=234#p234"><![CDATA[
Hi sorry for the confusion and please allow us to clarify.<br> <br> Data movements between CPU and GPU are performed by the GPU Copy Engine (CE).<br> However, Copy Engine activity is initiated by the CPU.<br> More specifically, copy commands are sent from the CPU to the GPU,<br> and this latter one fetches the command from a buffer and initiate the copy itself.<br> <br> More details are coming soon as we are about to release the model.<br> <br> Please let us know if you need further information.<br> <br> Nicola and Nacho<p>Statistics: Posted by <a href="http://localhost/memberlist.php?mode=viewprofile&amp;u=1594">Nacho_S</a> — Tue Mar 05, 2019</p><hr />
]]></content>
	</entry>
		<entry>
		<author><name><![CDATA[zero212]]></name></author>
		<updated>2019-03-04T15:02:35+01:00</updated>

		<published>2019-03-04T15:02:35+01:00</published>
		<id>http://localhost/viewtopic.php?t=124&amp;p=233#p233</id>
		<link href="http://localhost/viewtopic.php?t=124&amp;p=233#p233"/>
		<title type="html"><![CDATA[Re: Description of the WATERS Industrial Challenge 2019]]></title>

		
		<content type="html" xml:base="http://localhost/viewtopic.php?t=124&amp;p=233#p233"><![CDATA[
Hello everyone<br><br>I'm with a team currently working on the WATERS 2019 challenge.<br><br>Are the copy operations performed by the GPU or the CPU? Page 3 first sentence " The GPU integrates two major hardware components that are the responsible of the different GPU functions, namely an Execution Engine (EE) that is responsible for performing the parallel workload execution and a Copy Engine (CE) that is responsible of high bandwidth memory transfers." vs page 4 " A program running on a GPU works in the following way; when a program starts, the <strong class="text-strong">CPU</strong> copies the data and instructions from host memory domain to GPU memory, this copy is performed using the Copy Engine (CE) as seen in Figure 5." is a little bit confusing. <br><br>Would it be possible for us to get a model for the challenge? We can derive the hardware properties of the CPU and GPU from the paper. However  we are lacking the information in software properties, we can assume the ticks for the CPUs and other entities but it would be nice to have an comparable model. <br><br>Thank you in advance<p>Statistics: Posted by <a href="http://localhost/memberlist.php?mode=viewprofile&amp;u=25212">zero212</a> — Mon Mar 04, 2019</p><hr />
]]></content>
	</entry>
		<entry>
		<author><name><![CDATA[arne.hamann]]></name></author>
		<updated>2019-03-25T12:49:39+01:00 </updated>

		<published>2019-02-19T16:31:03+01:00</published>
		<id>http://localhost/viewtopic.php?t=124&amp;p=232#p232</id>
		<link href="http://localhost/viewtopic.php?t=124&amp;p=232#p232"/>
		<title type="html"><![CDATA[Description of the WATERS Industrial Challenge 2019]]></title>

		
		<content type="html" xml:base="http://localhost/viewtopic.php?t=124&amp;p=232#p232"><![CDATA[
<span style="font-size: 120%; line-height: 116%;"><span style="color:#FF0000">Update !!! : Amalthea model of the challenge now available in the following <a href="https://www.ecrts.org/forum/viewtopic.php?f=43&amp;t=126&amp;sid=8cc2f7c06f29e09b25476b714d191b3d" class="postlink">thread </a></span></span><br><br>I am glad to announce the description of the WATERS Industrial Challenge 2019 that is proposed by Bosch in cooperation with University of Modena.<br><br>Please note that the Amalthea performance file that is necessary to respond to the challenge will be delivered in 2-3 weeks.<br><br><strong class="text-strong">Abstract:</strong><br><br>Automotive  E/E architectures  are  currently  undergoing  a  radical  shift  in  the  way  they  are  designed, implemented  and  deployed.  Especially,  the  computational  power  and  communication  bandwidth required for new functionalities, such as automated driving or connected vehicle <br>functions (e.g. path planning,  object  recognition,  predictive  cruise  control),  exceed  the  capabilities  of  current  compute nodes  (mainly  micro-controller  SoCs)  and  is  leading  to  a  reorganization  of  automotive  systems  following  the  paradigm  of  so-called  centralized  E/E  architectures  that  are  based  on  a  new  class  of computing nodes featuring more powerful micro-processors and accelerators such as GPUs.<br>One  consequence  of  these  centralized  E/E  architectures  is  that  heterogeneous  applications  will  be co-existing on  the  same  HW  platform,  heterogeneous  not  only  in  their  model  of  computation (ranging  from  classical  periodic  control  over  event-based  planning  to  stream-based  perception applications)  but  also  in  their  criticality, in  terms  of  real-time  and  safety  requirements.  Ultimately, the  burden  of  integration  is  shifted  from  the  network  to  the  ECU  level  and  in  this  regard  typically from the vehicle manufacturer to the supplier of the control unit. <br><br>In order to cope with design and integration challenge, expressive performance models capturing the heterogeneity  of  the  hardware-software system  are  needed.  The  WATERS  industrial  challenge addresses  one  example  in  this  context,  where  the processing  power  offered  by  GPUs  and  their capability to execute parallel workloads is exploited to execute and accelerate applications related to advanced driver assistance systems.  <br>Thereby,  we  start  with  relatively  simple  models  that are  from  our  point-of-view,  on the  one  hand, sufficient to derive sensible performance predictions for the presented use-case, and that are, on the  other hand, not too far away  from established models used in the real-time community, or in other  words, there is a chance for existing models to be extended for addressing this challenge.<br>This document contains the following information: <ol style="list-style-type:decimal"><li>A primer on NVIDIA TX2 platform with focus on CPU–GPU interactions, scheduling, memory model, and offloading mechanisms as basis for performance modeling in Amalthea</li><li>A brief description of the challenge questions</li><li>A description of the considered application (a concrete Amalthea performance model will be provide to a later point in time)</li><li>A description of utilized Amalthea performance modeling approaches suitable to description  the software and hardware parts including their interaction for the given challenge.</li></ol>This description is a very good start for groups that want to work on the WATERS Industrial Challenge  2019.  However, we expect  that many  additional  questions will  arise  in  the  next  months.  Please  use this forum to  ask  those  additional  questions.  We  will  monitor  the  forum  and <br>answer your questions in a timely fashion.<dl class="file"><dt><span class="imageset icon_topic_attach"></span> <a class="postlink" href="http://localhost/download/file.php?id=100&amp;sid=3a00d0a38553769259dd6d3e8c889f13">WATERS_Industrial_Challenge_2019_final.pdf</a></dt></dl><br /><dl class="file"><dt><span class="imageset icon_topic_attach"></span> <a class="postlink" href="http://localhost/download/file.php?id=99&amp;sid=3a00d0a38553769259dd6d3e8c889f13">WATERS_Industrial_Challenge_2019.pdf</a></dt></dl><p>Statistics: Posted by <a href="http://localhost/memberlist.php?mode=viewprofile&amp;u=708">arne.hamann</a> — Tue Feb 19, 2019</p><hr />
]]></content>
	</entry>
	</feed>
