Overall we are pretty happy with the SGX, but our company is having a couple issues with drivers, and are curious to see if anyone else is having similar troubles.
We started development with the 3530 about a year ago, using a Mistral board, and are now on our own hardware (for the last… 6 months?). The company that provided our CPU boards (Cogent) went with Linux kernel 2.6.29 (because of the very nice Beagleboard mailing list, etc…), so we had to compile our own kernel modules. Everything seemed to work fine, until recently when we started running the cpu/kernel under heavier loads. Now it’ll run for several hours, and then our app gets stuck in sleeping, and the SGX ISR isn’t getting incremented.
We have done some custom kernel work, so this could definately be our fault, but we were wondering if this is a problem you have seen elsewhere?
Also, we still have screen tearing issues if something is moving rapidly – not a huge issue in our application, but we thought this was supposed to have been fixed a few releases ago? We are running the latest version of the drivers that we know about … .09
Do the debug mode drivers print anything interesting ? Can you describe what you mean by heavy load ?
Sorry about taking so long to reply, been really busy (and the bug stopped happening, but now it’s back…). We don’t get much debug output – it doesn’t happen on my machine which has the debug drivers, and it just basically hangs on the production machines – it looks like our application is waiting for the graphics to finish, and the SGX ISR stops incrementing in /proc/interrupts
We got it to stop the first time by slowing down the Linux Kernel jiffy rate, but now it’s back. (One of our engineers had turned up the jiffy rate to waaaay too fast…)
Are there release drivers out now, or are they still in Beta? I’ve been looking in vane for the web page that has the latest driver from TI…
Fixes continue to be made, and we are moving to 2.6.32 (an update will be available next week).
Is there a specific sequence of GL calls that causes this lock up ?
Well, we haven’t found a reproducible sequence of calls that causes it – we leave it running overnight with the same list of calls running and it usually dies after several hours. I think it’s our custom device drivers personally, but the boss wanted me to ask here if people have seen anything.
And a new rev of drivers would be very welcome! Would love to see preliminary release notes if they are available. 
Oh, and is this still a beta release, or?
Your product looks good, thanks for posting.
