LOGIN TO YOUR ACCOUNT

Username
Password
Remember Me
Or use your Academic/Social account:

CREATE AN ACCOUNT

Or use your Academic/Social account:

Congratulations!

You have just completed your registration at OpenAire.

Before you can login to the site, you will need to activate your account. An e-mail will be sent to you with the proper instructions.

Important!

Please note that this site is currently undergoing Beta testing.
Any new content you create is not guaranteed to be present to the final version of the site upon release.

Thank you for your patience,
OpenAire Dev Team.

Close This Message

CREATE AN ACCOUNT

Name:
Username:
Password:
Verify Password:
E-mail:
Verify E-mail:
*All Fields Are Required.
Please Verify You Are Human:
fbtwitterlinkedinvimeoflicker grey 14rssslideshare1
Bird, Robert F.
Languages: English
Types: Doctoral thesis
Subjects: QA76
Legacy code performance has failed to keep up with that of modern hardware. Many new hardware features remain under-utilised, with the majority of code bases still unable to make use of accelerated or heterogeneous architectures. Code maintainers now accept that they can no longer rely solely on hardware improvements to drive code performance, and that changes at the software engineering level need to be made.\ud \ud The principal focus of the work presented in this thesis is an analysis of the changes legacy Inertial Confinement Fusion (ICF) codes need to make in order to efficiently use current and future parallel architectures. We discuss the process of developing a performance model, and demonstrate the ability of such a model to make accurate predictions about code performance for code variants on a range of architectures. We build on the knowledge gained from such a process, and examine how Particle-in-Cell (PIC) codes must change in order to move towards the required levels of portable and future-proof performance needed to leverage the capabilities of modern hardware. As part of this investigation, we present an OpenCL port of the legacy code EPOCH, as well as a fully featured mini-app representing EPOCH. Finally, as a direct consequence of these investigations, we directly apply these performance optimisations to the production version EPOCH, culminating in a speedup of over 2x for the core algorithm\ud
  • No references.
  • No related research data.
  • No similar publications.

Share - Bookmark

Cite this article