387

I was wondering this earlier today, but thought that maybe no frames were complete. It's possible that there are 78 frames with 7/8 parts rendered so each is still missing one part. This is highly unlikely, for the frames should be calculated from the first frame to the last one. It could probably still happen if a lot of clients failed to render the parts in time. Each of the parts is calculated 3 times to make sure that no rendering errors and other anomalies get in. I performed a small routine check on the server and everything appears to be in order.

 hm, now I'm a little worried, could you see frames (finished or unfinished) on the server while perfoming the check?

 p.s.: here's some log from my client, which I suppose is rendering on this too:
So 18 Jul 2010 10:42:29 CEST|Open Rendering Environment|Reporting 2 completed tasks, requesting new tasks
So 18 Jul 2010 10:42:34 CEST|Open Rendering Environment|Scheduler request completed: got 0 new tasks
[b]So 18 Jul 2010 10:42:34 CEST|Open Rendering Environment|Message from server: No work sent[/b]
So 18 Jul 2010 11:09:20 CEST|Open Rendering Environment|Sending scheduler request: To fetch work.
So 18 Jul 2010 11:09:20 CEST|Open Rendering Environment|Requesting new tasks

don't know, if this is irregular..

The current session (387) really is awesomely heavy and probably should have been rendered in even smaller parts. Thus let's wait before jumping to conclusions.

I agree, no hurry, this isn't something important or time critical to me, more like a test of whats possible.

Next time I know better bout the parts, thanks for the information :)

but btw, my client now looks like this:
[url=http://i32.tinypic.com/2hi2pnq.jpg][/url]

If you check the advanced view, you can immediately see what's the actual cause for the error message. The unfortunate thing about the BOINC client that it displays (at least used to display) that silly error message even if there are simply no WU's to get from a project. We should actually probably contribute a change to that at some point in to the BOINC code.

 I couldn't really identify the cause in advanced view, it just stopped working on the project. I resetted the project, now it's working!

[quote]
This is where thit gets really hard to predict things, because some are calculated in just a couple of minutes basically.  
[/quote]

Why didnt I think of this before, surely the parts where just the shadeless background plane is, finish within no time, while the others render hours. This seems to me as a logical explanation why there are no finished frames yet.

NOW would be about the time to get worried. After going through some database entries I have discovered that most of the frames have 8/8 parts rendered. Total amount of finished frames is zero (the database entries are null). I believe we met something similar long ago, but I can't recall what was the cause. I'll see if I can get those frames created somehow. The rest of the crew is on holidays and I'm leaving on Thursday, so let's hope that we get this solved before that.

I've traced the error message, which seems to state that the parts received are inconsistent with the dimensions that are about to be "glued" together. There are currently two bug tests running. Both of them seem to produce frames correctly. Most mysterious. This problem might have to wait to Assembly where we'll ask jbk directly. We're sorry for the inconvenience. Once again, meet us at Assembly for candy compensation...