<div class="socmaildefaultfont" dir="ltr" style="font-family:Arial;font-size:10.5pt" ><div dir="ltr" style="font-family:Arial;font-size:10.5pt" ><div dir="ltr" >Jason,</div>
<div dir="ltr" > </div>
<div dir="ltr" >I concur with your decision since OMP and StreamExecutor fundamentally differ in how dependences between consecutive tasks are expressed. OMP uses task dependences to express constraint ordering between tasks that execute on the host and/or on a particular device. Obviously, a stream is a DAG but with very specific constraints (one linear ordering per stream), whereas DAG generated by OMP dependences are arbitrary DAGs. This is not a jugement statement, as in many ways stream are much more friendly to GPUs, it is just a decision that the OMP and StreamExecutor "language experts" settled on a different language expressivity/efficiency data point.</div>
<div dir="ltr" > </div>
<div dir="ltr" >I read your blog on the similarities and differences with great interest. I may venture to add another overlooked difference: OMP maps objects with references counts (e.g. first time an object is mapped, its ref count is zero, and the alloc on device and memory copy will occur; further nested map will not generate any alloc and/or communication). In summary, OMP primarily uses a dictionary of mapped variables to manage allocation and data transfer, whereas StreamExecutor it appears to explicitly allocate and move data.</div>
<div dir="ltr" style="font-family:Arial;font-size:10.5pt" ><div dir="ltr" style="font-family:Arial;font-size:10.5pt" ><div dir="ltr" style="font-family:Arial;font-size:10.5pt" ><div dir="ltr" style="font-family:Arial;font-size:10.5pt" ><div dir="ltr" > </div>
<div dir="ltr" >Thanks for your work on this, much appreciated</div>
<div dir="ltr" ><br>Alexandre<br><br>-----------------------------------------------------------------------------------------------------<br><span style="color:#0000CD;" >Alexandre Eichenberger, Master Inventor, Advanced Compiler Technologies</span><br><span style="color:#0000CD;" >- research</span>: compiler optimization (OpenMP, multithreading, SIMD)<br><span style="color:#0000CD;" >- info:</span> alexe@us.ibm.com http://www.research.ibm.com/people/a/alexe<br><span style="color:#0000CD;" >- phone</span>: 914-945-1812 (work) 914-312-3618 (cell)</div></div></div></div></div>
<div dir="ltr" > </div>
<div dir="ltr" > </div>
<blockquote data-history-content-modified="1" dir="ltr" style="border-left:solid #aaaaaa 2px; margin-left:5px; padding-left:5px; direction:ltr; margin-right:0px" >----- Original message -----<br>From: Jason Henline via Openmp-dev <openmp-dev@lists.llvm.org><br>Sent by: "Openmp-dev" <openmp-dev-bounces@lists.llvm.org><br>To: Andrey Bokhanko <andreybokhanko@gmail.com>, Chandler Carruth <chandlerc@google.com><br>Cc: llvm-dev <llvm-dev@lists.llvm.org>, cfe-dev <cfe-dev@lists.llvm.org>, "openmp-dev@lists.llvm.org" <openmp-dev@lists.llvm.org><br>Subject: Re: [Openmp-dev] [cfe-dev] RFC: Proposing an LLVM subproject for parallelism runtime and support libraries<br>Date: Mon, Mar 28, 2016 12:38 PM<br>
<div dir="ltr" ><div>I did a more thorough read through liboffload and wrote up a more detailed doc describing how StreamExecutor platforms relate to libomptarget RTL interfaces. The doc also describes why the lack of support for streams in libomptarget makes it impossible to implement some of the most important StreamExecutor platforms in terms of libomptarget (<a href="https://github.com/henline/streamexecutordoc/blob/master/se_and_openmp.rst" target="_blank" >https://github.com/henline/streamexecutordoc/blob/master/se_and_openmp.rst</a>). When I was originally optimistic about using liboffload to implement StreamExecutor platforms, I was not aware of this issue with streams. Thanks to Carlo Bertolli for bringing this to my attention.</div>
<div> </div>
<div>After having looked in detail at the liboffload code, it sounds like the best thing to do at this point is to keep StreamExecutor and liboffload separate, but to leave the door open to implement future StreamExecutor platforms in terms of liboffload. From the recent messages on this subject from Carlo and Andrey it seems like there is a general consensus on this, so I would like to move forward with the StreamExecutor project in this spirit.</div></div>
<div><div dir="ltr" >On Tue, Mar 15, 2016 at 5:09 PM Jason Henline <<a href="mailto:jhen@google.com" target="_blank" >jhen@google.com</a>> wrote:</div>
<blockquote style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex" ><div dir="ltr" >I created a GitHub repo that contains the documentation I have been creating for StreamExecutor. <a href="https://github.com/henline/streamexecutordoc" target="_blank" >https://github.com/henline/streamexecutordoc</a>
<div> </div>
<div>It contains the design docs from the original email in this thread, and it contains a new doc I just made that gives a more detailed sketch of the StreamExecutor platform plugin interface. This shows which methods must be implemented to support a new platform in StreamExecutor, or to provide a new implementation for an existing platform (e.g. using liboffload to implement the CUDA platform).</div>
<div> </div>
<div>I wrote up this doc in response to a lot of good questions I am getting about the details of how StreamExecutor might work with the code OpenMP already has in place.</div>
<div> </div>
<div>Best Regards,</div>
<div>-Jason</div></div>
<div><div dir="ltr" >On Tue, Mar 15, 2016 at 12:28 PM Andrey Bokhanko <<a href="mailto:andreybokhanko@gmail.com" target="_blank" >andreybokhanko@gmail.com</a>> wrote:</div>
<blockquote style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex" ><div dir="ltr" >Hola Chandler,</div>
<div dir="ltr" ><div> </div>
<div>On Tue, Mar 15, 2016 at 1:44 PM, Chandler Carruth via Openmp-dev <span dir="ltr" ><<a href="mailto:openmp-dev@lists.llvm.org" target="_blank" >openmp-dev@lists.llvm.org</a>></span> wrote:</div></div>
<div dir="ltr" ><div><div><blockquote style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex" ><div dir="ltr" ><div><div dir="ltr" ><span>It seems like if the OpenMP folks want to add a liboffload plugin to StreamExecutor, that would be an awesome additional platform, but I don't see why we need to force the coupling here.</span></div>
<div> </div></div></div></blockquote>
<div> </div></div></div></div>
<div dir="ltr" ><div><div><div>Let me give you a reason: while user-facing sides of StreamExecutor and OpenMP are quite different (and each warrants its place under the sun!), internal SE's offloading interface and liboffload are doing exactly the same thing. Why we want to duplicate code? As previous replies demonstrated, SE can't serve OpenMP's needs, while liboffload API seems to be general enough to serve SE well (though this has to be verified, of course -- as I understand, Jason is going to do this).</div>
<div> </div>
<div>Sure, there is no "must have need" to couple SE and liboffload, but this sounds like a solid software engineering decision to me. Or, quoting Jason, who said this much better than me:</div></div></div></div>
<div dir="ltr" ><div><div><div> </div>
<div><span style="color:rgb(0,0,0);white-space:pre-wrap" >> Although OpenMP and StreamExecutor support different programming models,</span></div>
<div><span style="color:rgb(0,0,0);white-space:pre-wrap" >> some of the work they perform under the hood will likely be very similar.</span></div>
<div><span style="color:rgb(0,0,0);white-space:pre-wrap" >> By sharing code and domain expertise, both projects will be improved and</span></div>
<div><span style="color:rgb(0,0,0);white-space:pre-wrap" >> strengthened as their capabilities are expanded. The StreamExecutor</span></div>
<div><span style="color:rgb(0,0,0);white-space:pre-wrap" >> community looks forward to much collaboration and discussion with OpenMP</span></div>
<div><span style="color:rgb(0,0,0);white-space:pre-wrap" >> about the best places and ways to cooperate.</span></div>
<div> </div></div></div></div>
<div dir="ltr" ><div><div><div><font color="#000000" ><span style="white-space:pre-wrap" >Espere veure't demà !</span></font></div>
<div> </div>
<div><font color="#000000" ><span style="white-space:pre-wrap" >Yours,</span></font></div>
<div><font color="#000000" ><span style="white-space:pre-wrap" >Andrey</span></font></div>
<div><font color="#000000" ><span style="white-space:pre-wrap" >=====</span></font></div>
<div><font color="#000000" ><span style="white-space:pre-wrap" >Enginyer de Software</span></font></div>
<div><font color="#000000" ><span style="white-space:pre-wrap" >Intel Compiler Team</span></font></div>
<div> </div></div></div></div></blockquote></div></blockquote></div>
<div><font face="Default Monospace,Courier New,Courier,monospace" size="2" >_______________________________________________<br>Openmp-dev mailing list<br>Openmp-dev@lists.llvm.org<br><a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/openmp-dev" target="_blank" >http://lists.llvm.org/cgi-bin/mailman/listinfo/openmp-dev</a></font></div></blockquote>
<div dir="ltr" > </div></div></div><BR>