High Performance Computing by Unknown

Author:Unknown
Language: eng
Format: epub
ISBN: 9783030507435
Publisher: Springer International Publishing

Details on the performance analysis of our initial implementation are provided in the next section. After profiling, we implemented some additional optimizations.

We eliminated some small host-to-device (“HtoD”) data transfers by using the DECLARE TARGET construct to declare some of the constants on the device. Furthermore, we declared some subroutine arguments to have the VALUE attribute. We found that the xlf compiler did not transfer the associated data as part of the kernel launch if they were passed by reference. However, passing the data by value circumvented the issue.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.