I'm trying to run a simulation over backburner and I keep getting this error on the BB machine:

Quote:2016/08/01 11:44:20 INF Job 'fumesky10' received and ready
2016/08/01 11:44:20 INF Launching 'Autodesk 3ds Max 2016 (64-bit)'
2016/08/01 11:45:00 INF 3dsmax adapter : Executing command line : C:\Program Files\Autodesk\3ds Max 2016\3dsmax.exe "-_pipe:1490384406" -server "C:\Users\Shawn\AppData\Local\backburner\ServerJob\fumesky10.max".
2016/08/01 11:45:00 INF [ServerMessage]ProcessStarted processId=6488
2016/08/01 11:45:00 INF Adapter returned unexpected code -2 for "C:\Program Files\Autodesk\3ds Max 2016\maxadapter.adp.exe" "-o" "LaunchApp" "-l" "Info" "-j" "C:\Users\Shawn\AppData\Local\backburner\ServerJob\fumesky10.xml.details" "-a" "C:\Users\Shawn\AppData\Local\backburner\ServerJob\fumesky10.zip"
2016/08/01 11:45:00 ERR Application load error: 3dsmax adapter error : 3dsmax.exe process no response.

2016/08/01 11:45:00 ERR 3dsmax adapter error : 3dsmax.exe process no response.

2016/08/01 11:45:30 ERR Application unload timeout (0 minutes)
2016/08/01 11:45:41 INF Adapter returned unexpected code -2 for "C:\Program Files\Autodesk\3ds Max 2016\maxadapter.adp.exe" "-o" "ShutDown" "-l" "Info"
2016/08/01 11:45:41 ERR Unknown Error

Hi there,
I've left n-sim over the night using backburner to send it over the network. Everything was great but then the computer stopped the sim and restarted... After some digging, I found that the problem was - ERR 3dsmax adapter error : Ping task time out. - default value 600 minutes. The problem is that I already have 138 frames correctly computed (.f3d zip). When I go local and try to continue sim, it says that it found data at frame 60 (basically because the job was restarted!!).
Now, even trying to continue sim, fume fx "thinks" that it only have 60 frames computed, but I have 138. Is there any way to force it to just go to frame 138 and continue the sim??
Thanks for any info on this!

Hi Kresimir,

I believe we discussed this years ago and now that I'm back to working with detailed blast sims, I'm running into this issue again. It only seems to occur for these blasts when cubic or advanced advection is used + the expansion is animated down or burn rate is animated down. I've been tweaking everything I can think of: quality, iterations, sensitivity 0, stride, the help file suggestions, etc. The only way I can get rid of them is to go with a non-cubic sim which is inferior in detail and motion. Any ideas on how these can be resolved? I was thinking of even using effectors testing for 'no temp, no velocity' and then add noise to break them up - if that would even work. Here's a screen shot, however you'll recognize the features from the issue noted in the help file under: UI / Simulation (the brown smoke image with arrows in it). In the help file it talks about fast moving smoke however I notice they only show up when the smoke slows down. Thanks. Best regards, Brian

Hi Kresimir,

I thought I'd start a new thread from my previous called 'retime ram usage' since this is really a new topic. I am trying to resolve the fact that I've used all ram on this node simming a WT sim to get the detail needed, and now need to slow the sim. I realize that the normal procedure is first running WT and then RT, however the RT process requires more ram than the WT sim needed and I'd need entirely new hardware to pull this off since this node is maxed at 72GB of ram.

(I'm calling 'post processing' RT since my primary focus is to retime the sim).

Is there a way to add the capability to the RT process to preserve the WT channel data so that an RT can be done first and then the WT process run after?

This would allow us to add maximum detail as our ram would allow during the WT process 'after' the RT process and and not have to keep it to approximately 60% so that there is still ram available for the RT phase. (The number could be wrong - that's what a few tests showed on this node).

I did try cheating the process by first running an RT of a default sim, preserving all channels, then loading the RT cache into the default folder and running WT on it. What I got was this strange error which didn't make sense since the WT data was definitely recorded with the default sim (arrow drawn pointing to it) and something about 'non null pointer to 12'... I'm guessing my cheat didn't work.

In the RT channels to be saved, I noticed there is TEX. I was hoping that was the WT channel, however since this didn't work, I'm guessing it's only texture coordinates.

In any case, could this feature be added? To allow for this order of simming: Default > RT > WT ?? Would help a lot of users to save on ram requirements for these huge detailed sims that need to be slowed.

I'm trying to retime an explosion WT cache that required up close to the maximum ram to sim on this node. The WT cache has all the necessary channels for retiming both fire and smoke. The RT sim crashes and runs out of ram. Ram usage for the WT sim was about 68GB out of 72GB max. The RT sim runs up to the full 72GB and 3dsmax crashes. Does the RT process require more than the WT sim did to calculate? I also have 'minimize' and some exclude channels active for the RT. Running 3dsmax 2015 and Ffx 4.0. Thanks in advance.

I get the attached errors when rebooting Max. The project I'm working on, after applying FFX successfully to the scene, crashed. Now I get the attached error messages when rebooting Max. How can I correct this? I tried rebooting the computer, but no luck. Thanks.

I'm getting a frozen noise pattern down inside the fire of an explosion sim. I haven't seen this before version 4.0 and am trying to isolate it. I've tried many different combinations of render steps, jitter, Vray AA, etc. and am not able to remove it. This sim has a lot of detail, hence it's not due to step/voxel issues from a low resolution sim. I'm using version 4.0 on this box. I'm going to test further, however any ideas would be helpful. Here's a screenshot. Thanks!

Hi,

After upgrading from FumeFX 4.0 to 4.0.1 I get Maya crashes running FumeFX. This is the stack Trace from the command Line:

FumeFX: Frame: 1
Stack trace:
/lib64/libc.so.6(+0x35670) [0x7fbf6e77c670]
/usr/autodesk/maya2016/lib/libVoxelFlow.so(+0x4c8b6) [0x7fbe6e8998b6]
FumeFXDyna::calcOneFrame(int, bool)
FumeFXDyna::SimulationOperator::operator()(FumeFXDyna*)
SimThreadFun(void*)
/lib64/libpthread.so.0(+0x7dc5) [0x7fbf6eb0fdc5]
clone
it's followed by quite a few of these Lines

QPixmap: It is not safe to use pixmaps outside the GUI thread

and some of these:

QObject:etParent: Cannot set parent, new parent is in a different thread
QObject:tartTimer: QTimer can only be used with threads started with QThread
QPainter::begin: Paint device returned engine == 0, type: 2
QPainter:etPen: Painter not active

This happens in the most simple scene which is a new fumeFX container with a simple source.

I didn't have these issues with 4.0 and can't find the download for 4.0 on the files section. Would I be able to get a link for 4.0 and I will rollback to this version?

Thanks.
Juri

Is there a script I could use to run the post sim every time the default cache finishes simming?

This would save me a lot of time. Is it even possible?

Hi

Our studio works intensively with Renderman 20.9 and Maya 2016 and we've been testing OpenVDB examples and we get them to work.
But when we export any .vdb from fumeFX 4.x and bringing it back in we either get a small foggy cube or Maya crashes.

Has anyone gotten FumeFX .vdb working in RenderMan and how do I go about in getting it to work correctly?

Kind Regards
Deon Oosthuizen