Optimizing real-time performance on Intel CPUs¶
Collaboration
This tutorial is a collaborative effort between Intel and Canonical. All instructions and source code are provided and released under the 3-Clause BSD License.
In this tutorial we will run through a number of configuration options to improve the real-time performance of a system that uses Intel CPUs. Some of the things that will be discussed include:
An example C program is provided to demonstrate the possible optimization steps to increase temporal isolation between best effort and real-time workloads on Linux systems. This test application creates a synthetic real-time load, and captures statistics on its performance.
Required hardware¶
Any 13th and 14th generation Intel CPUs that support Intel Time Coordinated Computing (TCC) should work. See the list in the Real-time at the Edge Gold Deck (slide 41), or check on Intel ARK.
Ideally your system BIOS should have the Intel® TCC Mode setting, but this is not a hard requirement for following this tutorial.
The following system was used to develop and validate the results obtained in this tutorial:
CPU: Intel® Core™ i5-1350PE processors
RAM: 16GB
DISK: 128GB