• Learn More about Intel® Time Coordinated Computing (Intel® TCC)

Intel® TCC comes with a suite of measurement and analysis tools that allow you to plan and optimize the performance of your real-time system. These include the Real-Time Readiness Checker, which checks the status of the system to make sure that it's enabled for real-time applications and the measurements library. The APIs in the Measurement Library are lightweight and offer minimal runtime overhead and high-measurement precision. The results can also be exported to JSON and other analysis tools. In this video, we're going to show you several examples of using the Measurement Library.

Hard real-time systems make strong assumptions about everything, including the application, the system, the environment, and all these are statically known, a priori ahead of time. The Real-Time Readiness Checker will help you check that your system is real-time ready, and the Measurement Library will help you measure the real-time latencies in your software environment. The keystone of real-time systems is predictability, including latencies and their periodicity, their arrival times, and their deadlines.

This will allow the developer to find the weakest element that determines the worst-case execution time.

Here's a diagram of the steps that you'll need to go through for creating a real-time application.

Step 1: Use the Real-Time Readiness Checker to quickly check the BIOS in other systems settings that might be affecting real-time performance.

Step 2: Run the application.

Step 3: Measure and analyze the behavior of your application. Instrument your code with the Measurements Library to gather latency statistics and visualize them for your needs. Lean on the measurement monitoring sample to get started. Take advantage of the measurement analysis sample, providing an application that will contract deadline violations and make histograms for any application instrumented with the library.

Step 4: After measurement and analysis, you may find that you have performance bottlenecks. You can address these and improve performance using other Intel TCC Tools such as the the cache configurator, or the cache allocation library, or the data streams optimizer.

In this demo, we are going to show you the monitoring measurements and deadline violations sample. This can be found in the TCC documentation.

First, go over and start the monitoring system. Next, we're going to run the single measurement sample. Here we're going to run a workload called approximation, and we're going to do this 10 times emulating the outliers. Now, if we go back over to the monitoring program, we can see that the cycles have been completed and that two out of a ten, exceeded the maximum amount of allowable time, causing a deadline violation.

In the next example, we're going to run a series of measurements and record their latencies to a log file to review later. Let's see how this is done.

Let's first run to an example of running the workloads without the measurement instrumentation, and then in our second example we'll show you how to modify that command in order to write the latencies out to a file you can review later. Next, we're going to use a modified version of this command in order to write the output latencies to a log file and record them for a later review. Once the command is finished, the latency will be written to a file and we can use the cat command or any editor to review them. That concludes our demo section.

To find out more, I'd encourage you to visit the Intel IoT Real-Time Technologies page and understand the foundation of real-time tuning that Intel offers. Please read the published tuning guides for the supported processor families, and to get hands on experience with TCC, download the BSP and other basic software packages from the Intel Resource and Design Center. Finally, download and try the TCC Tools package from the Intel® Developer Zone.