OctaneRender™ Standalone 3.00 alpha 7

Sun Apr 10, 2016 9:37 pm

Impossible to reduce width and height of fixed panes under a given size, previously it was possible to reduce them to quite nothing.

You can remove panes if you don't want them.

Sun Apr 10, 2016 9:42 pm

itou31 wrote:Same behaviour on my side. OK with PMC, but when swicth to PathTracing, the memory increase a lot.
When restart the application, it seems OK. But if you start another version and then restart back to alpha 7, same memory issue on DL or PT kernel.

Do you know a way to reproduce the problem?

Btw, memory usage of PT, DL and IC kernels is higher than PMC since they run more samples in parallel, but it depends on the parallel samples settings (see the new features post).

Sun Apr 10, 2016 10:08 pm

So I got some fume sims into octane finally.

(Yay!) And now for the questions for haze I guess...

I have an alembic animation starting at frame 1000. When try to load my open VDB sequence I cant tell it to use frame numbers in the filename so it starts at 0 from what I can tell. And there is no cache offset, so I cannot make the VDB's and alembic to match.

There also seems to be an offset between the geo and the vdb. In octane I need to offset my vdb with what seems to be "half" of my voxel size. (I need to dig more into this to be able to say for sure.)

Lastly, the window for loading multiple vdb's is not resize-able, so I can't see the filenames, only half the path, so it is hard to tell if I have loaded it correctly.

Other than that it is pretty awesome work you made in a short amount of time guys!!

I will return with more feedback as I go along.

Sun Apr 10, 2016 10:21 pm

Thanks for the feedback gardeler
Keep it coming

gardeler wrote: I have an alembic animation starting at frame 1000. When try to load my open VDB sequence I cant tell it to use frame numbers in the filename so it starts at 0 from what I can tell. And there is no cache offset, so I cannot make the VDB's and alembic to match.

I've put this on the to-do list - hopefully we'll get to it soon.

gardeler wrote: There also seems to be an offset between the geo and the vdb. In octane I need to offset my vdb with what seems to be "half" of my voxel size. (I need to dig more into this to be able to say for sure.)

I'm not aware of a problem right now but I'll check it out.

gardeler wrote: Lastly, the window for loading multiple vdb's is not resize-able, so I can't see the filenames, only half the path, so it is hard to tell if I have loaded it correctly.

On the todo list with the file animation works above

Mon Apr 11, 2016 1:48 am

manalokos wrote:Found a bug,

When you load a workspace layout, render restarts.

EDIT
And also changes the render target resolution.

I can't reproduce the render restart. What I see is that the bar meter representing samples/px goes to 0 until the next update, it should be ok again after some time. Is that what you see? We will fix this behaviour in the next release.

By the render resolution lock we mean this guy:

: resolution_lock.png (810 Bytes) Viewed 3006 times

If this is enabled, it will restart after each layout change because the resolution actually does change.

cheers,
Thomas

Mon Apr 11, 2016 1:58 am

Testing network rendering. I don't get it.

My 2 network cards shows that they are running at 98% load. But the GPU clock speed is only 574Mhz (2xGTX 780). Temp is very low - 62 degrees and they usually run at 83 degrees when using Octane 2.25.
What is going on?

Mon Apr 11, 2016 2:06 am

Zay wrote:Testing network rendering. I don't get it.

My 2 network cards shows that they are running at 98% load. But the GPU clock speed is only 574Mhz (2xGTX 780). Temp is very low - 62 degrees and they usually run at 83 degrees when using Octane 2.25.
What is going on?

I don't know, have never seen that behaviour. Maybe restart the slave and check if the problem persists? Which graphics driver?

Mon Apr 11, 2016 2:16 am

noldo wrote:some problem with environment & visible environment in conjunction with film width & film IOR (if film width > 1, then all black!!!)

I cannot reproduce this behaviour here. Can you give me a simple scene where I can reproduce this?

thanks,
Thomas

Mon Apr 11, 2016 2:30 am

abstrax wrote:I don't know, have never seen that behaviour. Maybe restart the slave and check if the problem persists? Which graphics driver?

On both - win7 driver 353.62
But problem fixed by restarting the slave PC. Was just restarting the software which obviously was not enough.

Mon Apr 11, 2016 10:04 pm

abstrax wrote:
itou31 wrote:Same behaviour on my side. OK with PMC, but when swicth to PathTracing, the memory increase a lot.
When restart the application, it seems OK. But if you start another version and then restart back to alpha 7, same memory issue on DL or PT kernel.
Do you know a way to reproduce the problem?

Btw, memory usage of PT, DL and IC kernels is higher than PMC since they run more samples in parallel, but it depends on the parallel samples settings (see the new features post).

Hi,
Here a little video demonstrate the issue.
I can reproduce it only on some of my scenes (OK with alpha 6). Once it fails, it will fail with all scenes without restarting.

bug_bis.mp4: (30.6 MiB) Downloaded 2821 times