Skip to content

Latest commit

 

History

History
71 lines (46 loc) · 1.73 KB

performance.md

File metadata and controls

71 lines (46 loc) · 1.73 KB

PTP performances

This document describes results in timing synchronization performance on a setup with two ptp slaves and a ptp master. Data were collected from the ptp4l slave algorithm internal metrics and log files and therefore graphed.

For this measurement campaign two different slaves were used. The main difference between them is the availability of hardware timestamping on the ethernet phy:

  • RaspberryPi3 (software only timestamp)
  • Variscite imx7-var-som board (hardware timestamp enabled)

The graphs below represent, on a per second base, the following informations:

  • The master offset value tracks the measured offset from the master (in nanoseconds).
  • States of the clock servo: s0 is unlocked, s1 is clock step, and s2 is locked.
  • The freq value represents the frequency adjustment of the clock (in parts per billion, ppb).
  • The path delay value represents the estimated delay of the synchronization messages sent from the master (in nanoseconds).

Test 1: PC as ptp4l master, rpi3 and imx7-var-som as ptp4l slaves

testbench

test-bench

The below pictures show the graph of the output data of ptp4l of 15 hours run.

RaspberryPi3

test-bench

Calculated offsets:

rms min max
21.5us -448us 797us

RaspberryPi3 (4 hours focus)

test-bench

Calculated offsets:

rms min max
16.5us -99us 84us

imx7

test-bench

Calculated offsets:

rms min max
12.3us -790us 790us

imx7 (4 hours focus)

test-bench

Calculated offsets:

rms min max
3.2us -34us 43us