AnsweredAssumed Answered

PTP (IEEE1588) Performance of BF518 or BF609 running Linux

Question asked by ErykD on Oct 15, 2012
Latest reply on Oct 16, 2012 by Aaronwu

Good day All,

 

I have been talking to Jiang Wu who was kind to answer all my questions to the extent possible. I know of the impressive results obtained by him and and Robert Peloquin running VDSP:

http://www.analog.com/library/analogDialogue/archives/43-11/IEEE_1588.html

 

Since our application requires Linux and we also need to obtain such good performance on BF518 or BF609 I am posting my question on EZone.

 

At the moment we have the BF518 evaluation kit and my tests using the Blackfin uClinux distribution (2010R1-RC5) are not promising.

With a direct copper connection between ADZS-BF518F-EZ-BRD and a grandmaster running of a OCXO (grandmaster is capable of synchronization down to 10ns) I am obtaining synchronization on the order of few microseconds with the hardware time-stamping support flags enabled.

Also in the documentation: http://docs.blackfin.uclinux.org/doku.php?id=uclinux-dist:ptp the shown results are far away from the VDSP solution.

According to this thread it seems that the state of the hardware support for time-stamping is questionable:

http://blackfin.uclinux.org/gf/project/uclinux-dist/tracker/?action=TrackerItemEdit&tracker_item_id=4642

 

Has anybody been able to obtain synchronization close to the BF518/BF609 datasheet spec on a setup similar to ours on linux? (or any other?)

Does anybody know what the status of the hardware timestamping for PTPd on uClinux is? Is there any work being done to improve the performance? Are there any documented tests?

 

Thank you for your time and assistance,

Eryk D

Outcomes