Dear Amber developers,
We are pleased to announce the release of a SYCL version of Amber20 for Intel Data Center GPU Max Series. The current version of the code enables PME simulations with several important features including AMBER and CHARMM force fields, thermostats, barostats, and most NMR type restraints.
The code is available in the branch “amber20-sycl-cmake” of the Amber git repository. Additional information and instructions for installation and testing are detailed in the file “README_sycl.md” in the Amber source root directory of this branch.
The code has been tested using Intel oneAPI 2024.0 on Intel Max Series Data Center GPUs (GPU Max 1100 and GPU Max 1550). The SYCL version involves a lot of new code, and we are looking for feedback from Amber developers on any issues that arise. If you want to test the code, you can get access to Intel Data Center GPUs on the Intel Developer Cloud. See
https://www.intel.com/content/www/us/en/developer/tools/devcloud/services.html and send an email to Kittur Ganesh (kittur.ganesh@intel.com) to facilitate access.
Feature completeness with respect to the CUDA implementation of pmemd and upgrade to the current Amber version are under development. Future SYCL code releases will also focus on support for Intel Arc series GPUs and portability across Intel, AMD, and Nvidia hardware.
The Amber20 SYCL code has been developed and tested in a collaborative effort with Intel’s Guoquan Chen and support by several others at Intel.
Please give the code a spin and send feedback to Andy Goetz (agoetz.sdsc.edu) and Guoquan Chen (guoquan.chen.intel.com). Thanks!
All the best,
Andy
—
Dr. Andreas W. Goetz
Associate Research Scientist
San Diego Supercomputer Center
Tel: +1-858-822-4771
Email: agoetz.sdsc.edu
Web: www.awgoetz.de
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Dec 19 2023 - 14:00:04 PST
