[Rtai] Newbie questions

[Graeme Foot]

> > ...
> >
> > 2) The socket communications are somewhat slower than I hoped for.  
> > I'm only managing 4 to 5 requests per cycle (with thread 1 only taking 
> > up 3% of its time).  When thread 1 is loaded up to 50% (with a busy 
> > wait) then it can only process 1 or 2 requests per cycle.  I have tried 
> > a couple the optimizations recommended here 
> > (http://www.ibm.com/developerworks/linux/library/l-hisock.html) with 
> > little effect.  I am currently only disabling nagle, though I did try 
> > the other suggestions without having much effect (but also not really
> > knowing which values to try).
> >
> > ...
> 
>
> Maybe want to look into RTnet? http://www.rtnet.org/
>
> Chris


Hi,

Thanks for the link.  Looking at it I don't think it will help in my case.  What I have is:

               LAN
                |
+---------------+-----------------+
| GUI for motion control system   |
|           Windows PC            |
+---------------+-----------------+
                |
                |
+---------------+-----------------+
|     New motion controller       |
|    Linux / RTAI / ethercat PC   |
+---------------+-----------------+
                |
         EtherCat fieldbus


So my new linux / RTAI pc has two NIC's, one is dedicated to the EtherCat fieldbus using the ethercat network card driver and the other NIC needs to accept socket commands from a windows based application.

Although the link between the windows PC and the Linux PC will be dedicated, from what I can see from the RTnet documentation there will be no point using RTnet because windows will only be able to interact using non-realtime communications anyway.  (Please correct me if I'm wrong.)

Also, the windows PC does not require deterministic messaging to the Linux PC.  All it needs is a high throughput (I'm hoping for about 10-20 commands per ms) and the assurance that handling the communications does not kill the hard realtime performance.


I've been playing with my system a bit.  I've disabled all sockets code so now I only have a hard realtime thread writing statistical data to sheared memory (similar to latency check code) and a soft realtime thread waiting on a semaphore that is released by the first (hrt) thread at the end of its cycle so that it can read the data and printf it.

I'm still getting overruns.  So here are a couple of questions:

1) Am I correct that to be able to use rt functions I need to call rt_task_init for a thread (and it becomes soft realtime)?

2) Am I correct that with the Fifo scheduler rt threads (both soft and hard realtime) will not be pre-empted by other rt threads?

3) Should I be using an async server (from README.LINUX_SERVER) even for soft realtime rt threads?

4) I've enabled the rtai_smi module just in case, but it doesn't make any difference.  I have an Intel® Celeron® M ULV, 1 GHz cpu.  As far as I can tell it uses SMI.  Does anyone know if I should be using the rtai_smi module with this cpu?


Thanks,
Graeme.