No
Yes
View More
View Less
Working...
Close
OK
Cancel
Confirm
System Message
Delete
Schedule
An unknown error has occurred and your request could not be completed. Please contact support.
Scheduled
Wait Listed
Personal Calendar
Speaking
Conference Event
Meeting
Interest
Schedule TBD
Conflict Found
This session is already scheduled at another time. Would you like to...
Loading...
Please enter a maximum of {0} characters.
Please enter a maximum of {0} words.
must be 50 characters or less.
must be 40 characters or less.
Session Summary
We were unable to load the map image.
This has not yet been assigned to a map.
Search Catalog
Reply
Replies ()
Search
New Post
Microblog
Microblog Thread
Post Reply
Post
Your session timed out.
This web page is not optimized for viewing on a mobile device. Visit this site in a desktop browser to access the full set of features.
2017 GTC San Jose

S7640 - Porting C++ Applications to GPUs with OpenACC for Lattice Quantum Chromodynamics

Session Speakers
Session Description

We'll describe our experience with using OpenACC to port a C++ library to run GPUs, focusing in particular on the issue of deep copy. The C++ library, Grid, is developed for numerical lattice quantum chromodynamics (LQCD) simulations, and is highly optimized for Intel x86 and many-core architectures. Our goal is to port it to run on NVIDIA GPUs using OpenACC so that its main code structure can be preserved and minimal code changes are required. We'll describe the challenges encountered and share the lessons learned during the porting process. In particular, due to the heavy use of templated abstractions, it is challenging to use OpenACC to deal with the data movement between the CPU and the GPU due to the deep-copy issue. We'll demonstrate that NVIDIA's virtual unified memory provides essential support for our porting effort. We'll also present initial performance results on Kepler and Pascal GPUs.


Additional Session Information
Intermediate
Talk
Computational Physics Programming Languages
Government / National Labs
25 minutes
Session Schedule