Install ROCm Compute Profiler

Before you begin, verify that your system is supported. For more information, see ROCm Core SDK components.

For advanced workflows, source builds, or custom configurations, see Install ROCm Compute Profiler from source.

Install the ROCm Core SDK

ROCm Compute Profiler (rocprofiler-compute) is included with the ROCm Core SDK on Linux. For the most complete installation, we recommend that developers use the amdrocm-core-sdk meta package.

For instructions, see Install AMD ROCm. Use the selector panel on that page to view instructions appropriate for your system environment.

Install ROCm profilers on Linux

Alternatively, if you want to install ROCm Compute Profiler as part of the ROCm Profiler package (a subset of the ROCm Core SDK amdrocm-core-sdk) without additional ROCm libraries and tools, install the amdrocm-profiler package. This includes the ROCm profilers, dependencies, and base packages.

1. Complete the ROCm installation prerequisites to install dependencies and configure GPU access permissions.

2. Install the ROCm Profiler package that matches your desired ROCm version.

   Package manager

   On Linux, package names use the following format:

   amdrocm-profiler<rocm_version>
   

   <rocm_version> represents the ROCm Core SDK version to install. Omit this suffix to install the latest available version.

   For example, to install the latest ROCm Profiler package release for supported GPU architectures:

   Debian-based distros

   Use the following command on Ubuntu and other Debian-based Linux distributions to install ROCm profilers:

   sudo apt install amdrocm-profiler
   

   RHEL-based distros

   Use the following command on RHEL, Oracle Linux, and other RHEL-based Linux distributions to install ROCm profilers:

   sudo dnf install amdrocm-profiler
   

   SLES

   Use the following command on SLES to install ROCm profilers:

   sudo zypper install amdrocm-profiler
   

   pip

   Use the following commands to create and activate a Python virtual environment, then install ROCm with the [profiler] extra:

   # Create and activate a Python virtual environment.
   python3 -m venv .venv
   source .venv/bin/activate
   
   # Install ROCm and the profilers from the AMD package repository.
   python -m pip install --index-url https://repo.amd.com/rocm/whl-multi-arch/ "rocm[profiler]"
   

Install from source

For runtime configuration options and modulefile-based environment setup, see Install ROCm Compute Profiler from source:

• Configuring the environment for profiling — configure the ROCPROF environment variable to select a profiling backend.

• Execution using modulefiles — load ROCm Compute Profiler via Lmod modulefiles for shared multi-user installations.

• Configuration variables — CMake variables for custom install paths, Python dependency locations, and build options.

Install from the tarball

1. Download the rocprofiler-compute specific tarball for the latest release from ROCm/rocm-systems.

2. Untar the downloaded tarball and navigate to the rocprofiler-compute directory.

3. Follow the installation steps under Install from source.

Install a nightly build

The TheRock build system also publishes nightly builds for the ROCm Core SDK and its components, including ROCm Compute Profiler. See Nightly release status for details.