BNL Home

Intel Knights Landing (KNL) Hackathon 2018

All programming paradigms are welcome

The Computational Science Initiative at Brookhaven National Laboratory, in partnership with the HEP Center for Computational Excellence and the SOLLVE Exascale Software Technology project, will host the first KNL Hackathon on February 26–March 2, 2018. All current and potential users of the Intel Knights Landing-based systems are encouraged to submit an application to participate.


Intel's 2nd Generation Xeon Phi processors, code-named "Knights Landing" or KNL for short, offer longer vector processing units, many cores (64-72 cores per processor), and hyper threading (up to 4 threads per core) ability, providing a theoretical peak floating point performance of several teraflops per node. In addition, they are equipped with high-bandwidth MCDRAM memory that can either be used as a far cache or a fast memory. These features mean that in order to take full advantage of KNL's computing horsepower, applications need to be refactored or optimized accordingly.


Currently there are several large-scale computing systems in the US that are based on Intel's KNL processors, including Cori-2 at NERSC, Theta at Argonne Leadership Computing Facility and Stampede-2 at Texas Advanced Computing Center. The KNL Hackathon will offer current and potential users of these systems the opportunity to have 5-day intense hands-on mentoring with computing experts from industry, national labs, universities and OpenMP standardization committee. The goal of the hackathon is to help the participants port or optimize their codes to take advantage of different levels of parallelism and memory hierarchies, which will not only allow them to run more efficiently on KNL-based systems, but also potentially improve the performance of their codes on other CPU-based systems.


Each team will come with 3-6 developers who are familiar with the respective application codes. And each team will be assigned 1-2 mentors who are experienced with OpenMP programming, profiling and/or KNL optimization. The teams will work on the application codes that they bring under the guidance of the mentors for 5 days. At the end of each day, teams are expected to give short (5-10 minutes) daily updates.

To get an idea of what the Hackathon may be like, check out the article and the video about the GPU Hackathon hosted at BNL in June 2017.


Team Prerequisites

  • The participants will come in teams of 3-6 people.
  • Team Application submission is now closed.
  • Deadline to submit an application is January 12, 2018.
  • Only accepted teams can participate.
  • Teams are assumed to be familiar with their application codes, and are comfortable working with Unix-based programming environments.
  • No experience with OpenMP, vectorization or KNL is required. We will send links to related documents and tutorials for the accepted teams before the hackathon. We will also have introductory tutorials during the Hackathon. But preparation before the Hackathon is strongly recommended.

computering facility

Important Dates

Jan 12, 2018 Deadline to Submit Team Applications
Jan 19, 2018 Notification of Accepted Team Applications
Jan 19, 2018 Registration (for accepted teams) Opens
Jan 31, 2018 Registration Closes
Feb 26, 2018 Hackathon Starts

Event Information

Organizing Committee

  • Barbara Chapman
  • Martin Kong
  • Meifeng Lin
  • Kerstin Kleese van Dam


Click on the "Event & BNL Information" tab above to find accommodation information.

Note: This event falls under Exemption D (Formal classroom training held at Federal facilities, which does not exhibit indicia of a formal conference as outlined in the Conference/Event Exemption Request Form.) Participation is contingent on application acceptance.

Top of Page

Last Modified: February 23, 2018