Begin forwarded message:

From: "Bertoni, Colleen" <bertoni@anl.gov>

Subject: OpenMP workshop tomorrow,

Date: April 25, 2019 at 6:35:54 PM CDT

To: "Luo, Ye" <yeluo@anl.gov>, "Monsalve Diaz, Jose Manuel" <jmonsalvediaz@anl.gov>, "Kwack, JaeHyuk" <jkwack@anl.gov>, "Knight, Christopher J." <knightc@anl.gov>, "Osborn, James C." <osborn@alcf.anl.gov>, "Li, Ying" <yingli@anl.gov>, "Romero, Nichols A." <naromero@anl.gov>, "MacKie-Mason, Brian" <bmackiemason@anl.gov>, "Patel, Saumil Sudhir" <spatel@anl.gov>, "Williams, Timothy J." <zippy@anl.gov>, "Garcia Martinez, Marta" <mgarcia@alcf.anl.gov>, "Adams, Corey James" <corey.adams@anl.gov>, "Parker, Scott J." <sparker@alcf.anl.gov>, "Kumaran, Kalyan" <kumaran@anl.gov>, "Zheng, Huihuo" <huihuo.zheng@anl.gov>, "Loy, Raymond M." <rloy@alcf.anl.gov>, "Applencourt, Thomas" <tapplencourt@anl.gov>, "Harms, Kevin" <harms@alcf.anl.gov>, "Ghadar, Yasaman" <ghadar@anl.gov>, "Chunduri, Sudheer" <sudheer@anl.gov>, "Rizzi, Silvio H.R." <srizzi@alcf.anl.gov>, "Shudler, Sergei" <sshudler@anl.gov>, "Insley, Joseph A." <insley@anl.gov>, "Raskar, Sid" <sraskar@anl.gov>, "Fang, Jun" <fangj@anl.gov>, "Morozov, Vitali A." <morozov@alcf.anl.gov>, "Rajak, Pankaj" <prajak@anl.gov>

Hi everyone,

For the OpenMP workshop tomorrow (9am-~12:30pm, TCS #1404 and 1405), the tentative agenda is

OpenMP Introduction and Overview
Hands-on Setup
OpenMP 4.5 Device Details
Hands-on (Steps to targeting a GPU with OpenMP)
Break
Porting existing code from MPI, MPI+OpenMP, etc. to target GPUs with OpenMP
OpenMP vs. OpenACC
OpenMP 5.0 vs. OpenMP 4.5

Slides will be sent out tomorrow morning. For hands-on work, we have a  reservation with queue R.PERFORM on Cooley from 9am-12:30pm. To use the  reservation, you’ll submit jobs to Cooley with 

$ qsub -A {performance, catalyst, datascience, etc.} -q R.PERFORM -t 180 -n 1 -I 
 
The setup for running the examples is here: /soft/compilers/clang-ykt/README on Cooley. It involves modifying your .soft.cooley, and then setting PATH and LD_LIBRARY_PATH.

The  details of how to run on Cooley will be mentioned again tomorrow, so no need to do anything now unless you want to. It is also possible to run the examples on JLSE using the nodes  with GPUs. 

Thanks,
JaeHyuk, Ye, Jose, Colleen