Problems when running Epiphany tests

Hardware related problems and workarounds

Re: Problems when running Epiphany tests

Postby mtimms2 » Fri Aug 15, 2014 4:18 pm

I wonder if we have a batch with missing components? Can you compare to see if the unpopulated components are in the same place on your working and non working boards?

Maybe the 4A Walmart type adapters are not stable enough a supply? I might try a bench PSU.

I have not tried running at a higher temperature yet, but this seems non ideal long term.

I too think I would have problems with an RMA as I have already had to pay duty once on import to the UK.
mtimms2
 
Posts: 4
Joined: Fri Aug 15, 2014 12:53 pm

Re: Problems when running Epiphany tests

Postby ralphmcardell » Fri Aug 15, 2014 4:53 pm

@mtimms2:

As far as I can tell (just did a quick compare with the bad board and one of the others) my A101040s are the same.

You could give it a go with a different PSU - certainly seems that "your power supply is not up to the job" is the default response to most Parallella problems and seems to have worked for some.
However, I have tried 3 PSUs so far:
- a small PC PSU which offers 5V @ 12A as part of a mini-cluster (4 Parallellas, one network switch)
- the PC PSU as above but only powering the cluster network switch and a single Parallella
- a TDK Lambda 5V 10A PSU powering just one Parallella
(experimented with setting voltages to ~4.76V and ~5.2V as well as ~5.02V)
- and recently, via the barrel connector, one of the Adapteva supplied wall-wart PSUs
And it did not in my case make any difference. So YMMV as they say.

I agree that running at a high temperature is not ideal long term - and in my case tricky to arrange with 3 other boards being cooled to more sensible temperatures in the same case - which is why I gave up in the end. However it might be a useful experiment to see if you can coax better behaviour out of your Parallella when using the Epiphany cores.
ralphmcardell
 
Posts: 12
Joined: Mon Dec 17, 2012 3:25 am
Location: London UK

Re: Problems when running Epiphany tests

Postby mtimms2 » Thu Aug 21, 2014 10:14 pm

Thanks for the reply.

On your suggestion, I have disconnected the fan and run the Zync up to 68 degC for a few hours to ensure the board and epiphany chip are both warm and up to temperature.

Still same issue in that any example causes a lockup of the board.

I also notice that running e-reset causes a lockup.

I am running this command from:
/opt/adapteva/esdk.5.13.09.10/tools/host/bin/e-reset

It looks like I can not reset the epiphany chip. I think all of the examples reset the device before running the code, so there is perhaps a reset issue which causes all of the examples to fail and lock up.

Is this a problem in the FPGA code? How can I debug the issue?
mtimms2
 
Posts: 4
Joined: Fri Aug 15, 2014 12:53 pm

Re: Problems when running Epiphany tests

Postby ralphmcardell » Sun Aug 24, 2014 5:07 pm

@mtimms2,

Sounds as if you have a more consistently fubar-ed board than I do :(

My bad board would sometimes lock up on the e_rest test so I had a peek at what resetting the Epiphany entails in the library source and although at the time I only had a quick look I did notice it was more involved than you might imagine (well, as I imagined!).

I have got to the point now of stepping back and setting up a more robust development environment, which when I get it up and running I hope will allow me to investigate your "How can I debug the issue?" question in some depth as I too would like to get a better handle on where things are going wrong.

I seem to remember that there are other posts that touch on debugging in some form or other and will probably be trying to track them down myself in a while.

Regards

Ralph
ralphmcardell
 
Posts: 12
Joined: Mon Dec 17, 2012 3:25 am
Location: London UK

Re: Problems when running Epiphany tests

Postby rowan194 » Fri Jan 16, 2015 8:23 am

6 months later... wondering if you guys ever figured out what the problem was? I seem to have the same issue, or one that seems to be similar. Basically, my Parallella is guaranteed to crash pretty soon after I start the Epiphany test suite. The board will also sometimes crash when just idling at the shell prompt. Occasionally it has locked up right after boot, before I can log in.

Here's my thread:

viewtopic.php?f=50&t=2100
rowan194
 
Posts: 17
Joined: Wed Jan 14, 2015 1:02 pm

Previous

Return to Troubleshooting

Who is online

Users browsing this forum: No registered users and 9 guests

cron