Intermittent Lock-Ups

Hardware related problems and workarounds

Intermittent Lock-Ups

Postby tgrant » Thu Jun 04, 2015 3:08 am

I have implemented a ray tracing application for the Parallella Desktop. After setting up the scene and initializing the Epiphany cores, the host polls a flag in the 'leader' core which signals that rendering has completed. The Epiphany cores access the scene data and a pixel buffer in shared DRAM.

The application completes the majority of the time, but occasionally the Parallella locks up and I am forced to cycle the power. Sometimes the lock-ups occur almost immediately, and other times they occur several minutes into a long render (why these simple renderings are taking several minutes is another matter entirely...). The application can run happily for >30 minutes, or it might lock up after 1 second, so it's hard to tell what's going on.

I'm at a loss as to why these lock-ups are occurring. I'm wondering if it could be a temperature issue, or some kind of memory corruption. I'm not sure if it's possible for the Epiphany cores to interfere with the ARM cores. I've placed a sleep in the polling loop to reduce the system burden of the busy loop, and the host CPU doesn't seem to be under excessive load when the Epiphany program is doing its thing.

If anyone has any troubleshooting suggestions, I would love to hear them.
tgrant
 
Posts: 5
Joined: Mon May 11, 2015 8:35 pm

Re: Intermittent Lock-Ups

Postby aolofsson » Thu Jun 04, 2015 10:00 am

Which SD image are you using?

What is your setup?

99% of the time, lockups like this have been caused by overheating, or more recently by the thermal safety deamon rebooting the device at 70 deg junction.
User avatar
aolofsson
 
Posts: 1005
Joined: Tue Dec 11, 2012 6:59 pm
Location: Lexington, Massachusetts,USA

Re: Intermittent Lock-Ups

Postby tgrant » Thu Jun 04, 2015 6:59 pm

I'm using the latest headless image, and my Parallella has the large heatsink that covers both the Zynq and the Epiphany chip. I'm not using a fan. High temperature would seem to be an obvious suspect, but as I said before the application can run at full bore for minutes on end with no issue, or it might lock up almost immediately after starting. The host processor isn't under heavy load when the lock-ups occur, so I'm wondering what other mechanism could cause the system to hang. Perhaps some kind of memory contention issue? It definitely isn't rebooting -- I need to cycle the power in order to bring the system back.

Is there some way I can verify (or rule out) the thermal shutdown possibility?
tgrant
 
Posts: 5
Joined: Mon May 11, 2015 8:35 pm

Re: Intermittent Lock-Ups

Postby tgrant » Fri Jun 05, 2015 2:53 am

I salvaged a fan from an old video card and slapped it on top of the heatsink, and so far the results are very encouraging. I haven't completely put the Parallella through its paces, but I haven't experienced any lock-up since attaching the fan. If I experience any more lock-ups I'll report back here. But I'm feeling pretty optimistic at this point.
tgrant
 
Posts: 5
Joined: Mon May 11, 2015 8:35 pm


Return to Troubleshooting

Who is online

Users browsing this forum: No registered users and 14 guests

cron