Description: This presentation will introduce TAU's profiling, tracing, and debugging support with focus on performance data collection, analysis, and program performance optimization.

Presenter: John Linford, Paratools, Inc.

Date/Location: July 16, 2013, ERDC

Distribution Statement D. Distribution limited to the Department of Defense and U.S. DoD contractors only, for administrative or operational use, 16 July 2013.  Other requests for this document shall be referred to the High Performance Modernization Office, 10501 Furnace Road, Suite 101, Lorton, VA 22079.

Distribution Statement D. Distribution authorized to the Department of Defense and U.S. DoD contractors only, Administrative or Operational Use, 9 November 2021.  Other requests for this document shall be referred to the High Performance Computing Modernization Office, 3909 Halls Ferry Road, Vicksburg, MS 39180.

Description: In this presentation, we will provide an overview of Arm Forge, a cross-platform, integrated environment for debugging and optimizing parallel codes at any scale. We will provide hands-on demonstrations of how Arm Forge reduces development time, simplifies debugging, and eases application performance enhancement.

Presenter(s): Beau Paisley, Arm
Location: Webcast
Date & Time: January 11, 2022, 2:00p - 4:00p ET

Distribution Statement D. Distribution authorized to the Department of Defense and U.S. DoD contractors only, Administrative or Operational Use, 11 January 2022.  Other requests for this document shall be referred to the High Performance Computing Modernization Office, 3909 Halls Ferry Road, Vicksburg, MS 39180.

Description: The GNU Debugger (gdb) allows users to trace and modify the execution of programs written in C, C++, Fortran, and various other languages. It is available on HPCMP machines and, as a CLI utility, can be more convenient than solutions that require graphics forwarding. We will introduce the use of gdb to trace application execution, demonstrate how to observe and modify program state, and show some less well-known features that make debugging with gdb easier and more efficient. No prior experience with gdb is required.

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.

Description: This one day course is a comprehensive introduction to parallel debugging using Allinea DDT, a commercial debugger produced by Allinea Software that is widely available on HPCMP systems.

Presenter: Nathan Stone, Carnegie Mellon University

Date/Location: July 11, 2014, ERDC, Vicksburg MS

Distribution Statement D. Distribution limited to the Department of Defense and U.S. DoD contractors only, for administrative or operational use, 11 July 2014.  Other requests for this document shall be referred to the High Performance Modernization Office, 10501 Furnace Road, Suite 101, Lorton, VA 22079.

Distribution Statement D. Distribution authorized to the Department of Defense and U.S. DoD contractors only, Administrative or Operational Use, 28 January 2021.  Other requests for this document shall be referred to the High Performance Computing Modernization Office, 3909 Halls Ferry Road, Vicksburg, MS 39180.

Description: Deal.II (Differential Equation Analysis Library II) is a state-of-the-art finite element Toolbox focused on generality, parallelism, and extensibility. It provides a powerful software library that is amongst the largest of the libraries supporting numerical PDE solvers and may serve as a potential alternative to currently supported commercial and in-house finite-element libraries and tools on the DoD HPC.

This seminar will offer HPC users the opportunity to gain experience with open-source software management using Spack and CMake, from installation to testing/debugging, with guidance on managing software dependencies, optimizing HPC environment, and identifying compiler requirement as well as libraries.

CUI

Description: What are MPI and OpenMP? MPI: Distributed memory model on Distributed network, Message based, Flexible and expressive. OpenMP: Shared memory model on Multi-core processors, Directive based easier to program and debug. Message Passing Interface (MPI) is a library specification for message-passing, proposed as a standard by a broadly based committee of vendors, implementers, and users. - OpenMP is a specification for a set of compiler directives, library routines, and environment variables that can be used to specify shared memory parallelism in Fortran and C/C++ programs.will be presented.

Presenters: Dr. Ross Smith, Engility PETTT

Date/Location: June 6-7, 2018, ARL

DISTRIBUTION STATEMENT A. Approved for public release: distribution unlimited.

Description: Interactive HPC use involves running programs on HPCs in real time instead of submitting batch jobs. Users may need to iterate on shorter tasks, debug or optimize software, run GUIs that include a compute component, or run tasks like compiling software or compressing data that benefit from using many cores and should not be done on login nodes. Users may need direct and dynamic control of software running on one or more nodes. This training is split into two sessions.

Description: PETSc solvers are critical to the performance of many highly-parallel scientific and engineering applications. However, problem-specific and machine-specific factors strongly affect the performance of these solvers. Because there is no set of one-size-fits-all solver options for these applications, measuring and understanding the performance of PETSc routines within specific applications is an important step in reducing running time and making efficient use of compute power.

This seminar will describe, including examples, the tools built in to PETSc for logging and profiling applications.  These tools allow a detailed look inside PETSc-based applications without the need for an external tool such as a profiler or debugger.