Before you begin, make sure your development environment is set up. If you have not already configured your environment, go to
Configure Your System before proceeding.
You can use a terminal window or Visual Studio Code* for command-line development. Some tasks can be automated using extensions. To learn more, see
Using Visual Studio Code with Intel® oneAPI Toolkits.
To compile and run a sample:
- Use the oneAPI CLI Samples Browser to locate a sample project.
- Build and run the project using
Make* or CMake*.
Download Samples using the oneAPI CLI Samples Browser
Use the oneAPI CLI Samples Browser to explore the collection of online oneAPI samples. The oneAPI CLI Browser is a single-file, stand-alone tool that does not require any extra libraries. You can copy samples to your local disk as ready-to-build projects. The build instructions are included as part of the sample in a README file. Most oneAPI sample projects are built using Make or CMake.
Watch a video guide on creating a sample project with
Intel® oneAPI using the command line.
To download and set up a sample:
- Open a
terminal
window.
- If you did not complete the steps in Option 2: One time set up for setvars.sh in the
Configure Your System section,
set
system variables by
sourcing setvars. You can choose to use either the Unified Directory Layout or the
Component Directory Layout to ensure components version consistency. Learn more in the oneAPI Development Environment Setup.
The commands below assume you used the default install location. If you customized the installation folder, adjust the path to
setvars | oneapi-vars accordingly.
Component Directory Layout:
For system-wide installations (requires root or sudo privileges):
. /opt/intel/oneapi/setvars.sh
For local installations:
. ~/intel/oneapi/setvars.sh
Unified Directory Layout:
For system-wide installations (requires root or sudo privileges):
. /opt/intel/oneapi/<toolkit-version>/oneapi-vars.sh
For local installations:
. ~/intel/oneapi/<toolkit-version>/oneapi-vars.sh
- In the same
terminal
window, run the application (it should be in your PATH):
oneapi-cli
The oneAPI CLI menu appears:

- Use the up and down arrow keys to select
Create a project, then press
Enter
- The language selection will appear. For your first project, select
cpp, then press
Enter. The toolkit samples list appears.

- Select the
Vector Add sample.
Vector Add is a simple test application that help verify that the tools are setup correctly and can access your system's GPU:

After you select a sample, press
Enter.
- Specify the location for the project. The default location includes the path from where the CLI Browser was run and the name of the project.
Press
Tab to select Create, then press
Enter:

Build and Run a Sample Using Make
Note
Some samples require additional steps or arguments for building and/or running the sample. Review the sample's
README.md file for specific details regarding how to build and run the sample.
- Open a command prompt.
- Navigate to the directory which you specified when creating (downloading) the project.
- Configure the project using either the
buffer-based implementation or the
Unified Shared Memory (USM) based implementation
. Read more about these memory management techniques in the
Vector Add Sample page.
Buffer-based implementation:
mkdir build
cd build
cmake ..
Unified Shared Memory (USM) based implementation:
mkdir build
cd build
cmake .. -DUSM=1
- Build the program.
make cpu-gpu
- Run the program for Unified Shared Memory (USM) or buffers.
./vector-add-buffers
./vector-add-usm
- Optional: Clean the program.
make clean
Build and Run a Sample using CMake
Vector add can
only be run with Make. You can run a
different sample using CMake, where
<sample_name> is the name of the sample you want to run.
To find a sample that uses CMake, browse the
oneAPI Samples GitHub repository and view the
README file for each sample to see if the sample can be run with CMake.
- If necessary, re-run the command-line utility and select a CMake project that contains a
CMakeLists.txt file.
cd <sample_name>
- Navigate to the build directory.
mkdir build
cd build
- Build the program. Run CMake in the build directory to create the makefile. Use the newly created makefile to build the executable.
cmake ../.
make VERBOSE=1
Note
Some samples require additional steps or arguments for building and/or running the sample. Review the sample's
README.md file for specific details regarding how to build and run the sample.
- Run the program.
make run
- Clean the program.
make clean
Compile and run a sample for FPGA
You can run the
vector-add sample (or any FPGA SYCL code) in the following modes:
- Emulation: Verifies the code correctness. Compilation completes in few seconds. Use this mode if you are using the
Intel® oneAPI Base Toolkit
- Report: Generates a static optimization report for design analysis. Compilation can take a few minutes to complete. When completed, you can find the reports in
<project_name>.prj/reports/report.html. This can be used with the
Intel® oneAPI Base Toolkit. For more information about the reports, refer to the
FPGA Optimization Guide for Intel® oneAPI Toolkits.
- Hardware:
Generates the actual bitstream on an FPGA device. Compilation can take few hours to complete. Use this mode to measure performance. To use this mode, download the Intel® Quartus® Prime Pro Edition software and BSPs separately. For more information, refer to the
Install Software for Intel® FPGA Development Flow section in the
Intel® oneAPI Toolkits Installation Guide for Linux* OS and
Intel® FPGA Add-On for oneAPI Base Toolkit web page.
After downloading the
vector-add design example using the oneAPI CLI Samples Browser, perform the following steps to compile and run the design:
- Change to the directory containing the
vector-add design example using:
cd <vector-add directory on the same system>
- Run the following
make clean command before you start compiling:
make clean
- Compile the design using one of the following options:
Compile for emulation using:
make fpga_emu
Compile the report using:
make report
Note
You can view the report at
vector-add_report.prj/reports/report.html. This does not generate an executable.
Compile for hardware (takes longer to complete) using:
make fpga
Note
If you compiled the hardware on a development system, copy the executable file
vector-add.fpga to the runtime system.
- Run the design using one of the following options:
- Run the design for emulation using:
./vector-add-buffers.fpga_emu
./vector-add-usm.fpga_emu
Note
If you are using a separate development system, perform this step on that system.
- Run the design on FPGA hardware using:
./vector-add-buffers.fpga
./vector-add-usm.fpga