[etherlab-users] Xenomai performance
Christoph Schröder
christoph.schroeder at helmholtz-berlin.de
Thu Sep 29 12:56:01 CEST 2016
Hi Graeme,
Thanks for the answer. I have 4 modules (EK1100, EL1008, EL2004 and
EL3202) from which only the BusCoupler and the EL2004 (also used in
Xenomai example) are active.
Here are my delays:
$ ethercat slaves -v | grep delay
DC system time transmission delay: 0 ns
DC system time transmission delay: 145 ns
DC system time transmission delay: 285 ns
DC system time transmission delay: 440 ns
I tested the application over last night and had no losses, 12,5kHz
seems stable for my configuration. This means 80000ns cycle time so I
dont think the transmission delays are a factor here. The task itself is
not doing much, just inverting the output every other second + the added
rt_printf of the cycle_counter. It's basically the same as the xenomai
example in ./examples/xenomai of the master sources. I also set the task
to the maximum priority (99).
I attached the example with all my changes. If anyone uses Xenomai and
has an EL2004 he may if he can get better results. I will also test it
with Xenomai3 and a newer kernel (3.16) in the future.
Best regards,
Christoph
On 09/28/2016 11:24 PM, Graeme Foot wrote:
> Hi,
>
> I don't use xenomai so can't comment on that side, but just curious, how many ethercat modules do you have?
>
> Also, if you run the ethercat command:
> ethercat slaves -v | grep 'transmission delay'
>
> What is the delay time at the last module?
>
> This is how long the ethercat frame takes to reach that module. It will then also take the same amount of time to return (assuming a linear topology). You will also need to factor in processing time overhead of the master to prepare and send the frame and receive the frame, and also add the worst case jitter time. This is in theory the shortest turnaround time you can expect per frame. It does not factor in any processing time your application may need.
>
> Regards,
> Graeme.
>
>
> -----Original Message-----
> From: etherlab-users [mailto:etherlab-users-bounces at etherlab.org] On Behalf Of Christoph Schröder
> Sent: Thursday, 29 September 2016 3:25 a.m.
> To: etherlab-users at etherlab.org
> Subject: [etherlab-users] Xenomai performance
>
> Hi all,
>
> I am currently testing the EtherCAT master with Xenomai 2.6.5 on a Debian wheezy (ipipe patched kernel 3.2). Using the xenomai example from the master I can go up to 12,5kHz (maybe a bit more). I got lost frames from time to time if I go up to 15kHz and a lot of lost frames if I go up to 20kHz.
> This is a lot better than non-rt, but in the master documentation there was noted that they could go up to 25kHz on a 2GHz System with RTAI. My testsystem is a Core i5-4690 CPU @ 3.50GHz. I configured everything as described in the Xenomai installation instruction. Are my results reasonable resp. is RTAI simply better or should I expect some misconfiguration?
>
> My results from the xenomai latency test seem to be fine.
> $ sudo ./xeno latency -T 25
> == Sampling period: 100 us
> == Test mode: periodic user-mode task
> == All results in microseconds
> warming up...
> RTT| 00:00:01 (periodic user-mode task, 100 us period, priority 99)
> RTH|----lat min|----lat avg|----lat max|-overrun|---msw|---lat
> best|--lat worst
> RTD| -2.307| -2.152| 0.550| 0| 0| -2.307| 0.550
> RTD| -2.516| -2.160| 4.011| 0| 0| -2.516| 4.011
> RTD| -2.319| -2.170| 0.515| 0| 0| -2.516| 4.011
> RTD| -2.324| -2.165| -0.322| 0| 0| -2.516| 4.011
> RTD| -2.336| -2.127| 2.644| 0| 0| -2.516| 4.011
> RTD| -2.350| -2.146| 2.139| 0| 0| -2.516| 4.011
> RTD| -2.542| -2.160| -0.075| 0| 0| -2.542| 4.011
> RTD| -2.363| -2.146| 2.111| 0| 0| -2.542| 4.011
> RTD| -2.590| -2.156| 0.461| 0| 0| -2.590| 4.011
> RTD| -2.703| -2.158| -1.017| 0| 0| -2.703| 4.011
> RTD| -2.701| -2.157| 6.590| 0| 0| -2.703| 6.590
> RTD| -2.695| -2.156| -1.049| 0| 0| -2.703| 6.590
> RTD| -2.684| -2.158| 0.496| 0| 0| -2.703| 6.590
> RTD| -2.712| -2.154| -1.079| 0| 0| -2.712| 6.590
> RTD| -2.693| -2.158| 6.106| 0| 0| -2.712| 6.590
> RTD| -2.700| -2.159| -1.005| 0| 0| -2.712| 6.590
> RTD| -2.712| -2.156| -0.882| 0| 0| -2.712| 6.590
> RTD| -2.642| -2.159| -0.836| 0| 0| -2.712| 6.590
> RTD| -2.608| -2.152| 5.227| 0| 0| -2.712| 6.590
> RTD| -2.686| -2.158| -0.742| 0| 0| -2.712| 6.590
> RTD| -2.653| -2.158| 0.883| 0| 0| -2.712| 6.590
> RTT| 00:00:22 (periodic user-mode task, 100 us period, priority 99)
> RTH|----lat min|----lat avg|----lat max|-overrun|---msw|---lat
> best|--lat worst
> RTD| -2.723| -2.160| -1.258| 0| 0| -2.723| 6.590
> RTD| -2.643| -2.161| 0.059| 0| 0| -2.723| 6.590
> RTD| -2.366| -2.168| -0.458| 0| 0| -2.723| 6.590
> ---|-----------|-----------|-----------|--------|------|----------------
> ---|-----------|-----------|-----------|--------|------|---------
> RTS| -2.723| -2.156| 6.590| 0| 0| 00:00:25/00:00:25
>
> Thanks and best regards,
> Christoph
>
> ________________________________
>
> Helmholtz-Zentrum Berlin für Materialien und Energie GmbH
>
> Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.
>
> Aufsichtsrat: Vorsitzender Dr. Karl Eugen Huthmacher, stv. Vorsitzende Dr. Jutta Koch-Unterseher
> Geschäftsführung: Prof. Dr. Anke Rita Kaysser-Pyzalla, Thomas Frederking
>
> Sitz Berlin, AG Charlottenburg, 89 HRB 5583
>
> Postadresse:
> Hahn-Meitner-Platz 1
> D-14109 Berlin
>
> http://www.helmholtz-berlin.de
> _______________________________________________
> etherlab-users mailing list
> etherlab-users at etherlab.org
> http://lists.etherlab.org/mailman/listinfo/etherlab-users
________________________________
Helmholtz-Zentrum Berlin für Materialien und Energie GmbH
Mitglied der Hermann von Helmholtz-Gemeinschaft Deutscher Forschungszentren e.V.
Aufsichtsrat: Vorsitzender Dr. Karl Eugen Huthmacher, stv. Vorsitzende Dr. Jutta Koch-Unterseher
Geschäftsführung: Prof. Dr. Anke Rita Kaysser-Pyzalla, Thomas Frederking
Sitz Berlin, AG Charlottenburg, 89 HRB 5583
Postadresse:
Hahn-Meitner-Platz 1
D-14109 Berlin
http://www.helmholtz-berlin.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: main.c
Type: text/x-csrc
Size: 7817 bytes
Desc: not available
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20160929/271addd5/attachment-0005.c>
More information about the Etherlab-users
mailing list