Description: File input and output (I/O) often become a performance bottleneck for large-scale, parallel applications as I/O workloads and hardware have become increasingly complex on DoD HPC systems. To achieve high performance, data-intensive scientific applications use multiple layers within the I/O software stack comprising high-level I/O libraries, such as NetCDF and HDF5, and MPI I/O and leverage parallel file systems, such as the Lustre file systems available on HPCMP systems. Designing efficient parallel scientific applications therefore requires an understanding of the underlying I/O operations, their interaction with the supporting I/O libraries, and, ultimately, the file system upon which the data resides.
This seminar provides an overview and introduction to Darshan, a scalable I/O characterization tool that can be used to investigate and tune the I/O behavior of complex parallel applications to achieve efficient I/O performance. Users will also be provided with guidance on installing and using Darshan on HPCMP systems and analyzing Darshan output using sample data from the open-source I/O benchmarking application benchIO.
| Presenter(s): Dr. Harley Klein, GDIT / PET Location: Webinar Date & Time: April 23, 2026, 2:00p – 3:30p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
- Presenter: Harley Klein
Description: The CUDA-GDB debugger provides existing GDB debugging features for host (non-GPU) code and adds features to support debugging CUDA device code. It allows users to seamlessly debug both GPU and CPU code within the same application. For example, CUDA-GDB allows setting breakpoints, single-stepping CUDA applications, inspection and modification of program state for individual CUDA threads. This training will demonstrate the use of CUDA-GDB on an HPCMP computer beginning with basic commands for CUDA debugging and working up to examples of debugging small CUDA applications.
Brief reviews of core CUDA concepts and basic GDB commands will be given, but some prior experience with both tools may be beneficial for following some examples.
| Presenter(s): Dr. Nathan Bowman, GDIT / PET Location: Webcast Date & Time: December 11, 2:00p – 3:30p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
- Presenter: Nathaniel Bowman
Description: Discover the latest advancements in the TI-21 software development environment since the initial training sessions. Learn how to debug and optimize applications, including performance debugging on the TI-21 system. Harness the power of vendor-provided software, such as NVIDIA and Red Hat, within self-contained containers. Experience seamless access to various software, including operating systems not installed natively. Tailor your environment to specific versions of packages, libraries, or those not included natively on the system. Run applications in isolated environments and ensure secure access to software and data.
| Presenter(s): Dr. Bracy Elton - Penguin Solutions Location: Webcast Date & Time: September 18-19, 2024, 9:00a - 4:00p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
Raider Penguin Nautilus TI-21
Description: This training provides users with an overview and introduction to the Penguin Computing TI-21 “Raider” HPC located at the AFRL DSRC. Users will be shown system features, use cases, data management, running jobs, and best practices for productively using Raider.
- Introduction & System Features
- Typical End User Use Case
- Data Management & Movement
- Best Practices
High-performance computing plays a crucial role in many Department of Defense (DOD) projects, enabling scientists and researchers to achieve their science and technology objectives in various scientific disciplines, such as climate, weather, and ocean modeling, computational fluid dynamics, and computational chemistry.
By using complex applications on these specialized HPC systems, they can predict the behavior of engineered systems in complex environments and analyze patterns in physical systems, enabling scientific discoveries and improvements in warfighting capabilities in a much faster, more efficient manner.
| Presenter(s): U.S Airforce & Penguin Solutions Location: Webcast Date & Time: April 29-30, 2024, 9:00a - 3:00p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
Description: This seminar will help users develop a better understanding of HPC networks and how they can reduce communication costs to develop more efficient and scalable HPC codes. This seminar will include an overview of the networks on DoD HPC systems, discussion of key factors that impact performance, discussion of available tools for analyzing network performance, and optimization techniques for improving the network performance of HPC codes.
| Presenter(s): Dr. Paul Eller, GDIT/PET Location: Webcast Date & Time: August 25, 2022, 2:00p - 4:00p ET |
Controlled by: DoD HPCMP Controlled by: PET Program CUI Category: OPSEC Limited Dissemination Control: FEDCON POC: Mr. Ronald Hedgepeth, pet@hpc.mil |
CUI
- Presenter: Paul Eller
Description: The AFRL DSRC and the HPE team will be introducing users to the new Mustang supercomputer. Training will cover user access, commands, configuration along with in depth coverage of the Intel OmniPath Architecture network.
Presenters: Hewlett Packard Enterprise (HPE), Intel Corporation, and NVIDIA
Location: AFRL Bldg 676, Room 105, WPAFB, OH
Date & Time: January 8-11, 2019, 8:30a - 5:00p ET
Distribution Statement D. Distribution limited to the Department of Defense and U.S. DoD contractors only, for administrative or operational use, 11 January 2019. Other requests for this document shall be referred to the High Performance Computing Modernization Office, 3909 Halls Ferry Road, Vicksburg, MS 39180.