Column: Embedded design


The CPU affi nity of the interrupt and corresponding handler matches the CPU affi nity of the process receiving packets, with packet loss being minimised because there’s no time wasted moving data between cores


Figure 8: The top command shows that all four cores are being throttled to their maximum


Figure 9: A zoom view of the capture


CPU cores; zoom into the middle of the capture to look at the execution sequence of the eth0 interrupt handler and the iperf instance. Notice that the time between completion of the eth0 interrupt handler execution and the start of the iperf execution is now approximately 40 microseconds; see Figure 9.


While the number itself isn’t important


(although it is interesting that the time is less under load!), it is the same order of magnitude when the system is under load versus when it is not – 40 microseconds versus 55. This is a tremendous feature of the Linux kernel, where even with a userspace application seemingly hogging all four cores of


Figure 10: A view showing the CPU is indeed under heavy load


the system, it still ensures that other userspace applications are not starved for CPU resources and that intercore communication is not affected. If we take a step back, we can see all


the processes that represent the “stress” application, confirming that the CPU is indeed under heavy load; see Figure 10.


www.electronicsworld.co.uk April 2021 11


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44