OctaneRender® for Maya® beta 3.03f [OBSOLETE]
Moderator: JimStar
Hmmm, they worked fine on my system and installed into both 2012 and 2013 at the same time.Joss wrote:Tbh, neither of the latest installers was able to detect maya 2012 installation, when 2013 is installed along with 2012, jfyi.
T.
Win10 x64|i7-9750H 2.6 GHz|32 GB RAM | RTX2080 max Q 8GB
that is weird, the install was the same for 3.03f as it was for all the others for me.Joss wrote:Tbh, neither of the latest installers was able to detect maya 2012 installation, when 2013 is installed along with 2012, jfyi.
Maya 2013 / W7 64bit / Intel i7 980 / 24GB / Gigabyte gtx 580 3GB + 2*EVGA gtx 680 4gb / MB Gigabyte X58A-UD3R / NZXT Phantom - white/black case
Maya 2013/W7 64bit/Intel i7 3820/16GB/ MB Gigabyte GA-X79-UD3 /4*EVGA gtx 680 4gb/SSD 6Gb.s /CM HAF X case
Maya 2013/W7 64bit/Intel i7 3820/16GB/ MB Gigabyte GA-X79-UD3 /4*EVGA gtx 680 4gb/SSD 6Gb.s /CM HAF X case
1. Do you mean that if you install only 2012, all is detected OK?Joss wrote:Tbh, neither of the latest installers was able to detect maya 2012 installation, when 2013 is installed along with 2012, jfyi.
2. Is your 2012 installed to the default path or to something else?
3. Can you check if the registry key "HKLM\SOFTWARE\Autodesk\Maya\2012\Setup\InstallPath" exists and holds the right value?
Last edited by JimStar on Fri Nov 23, 2012 8:00 am, edited 1 time in total.
It is the engine issue, I can't do anything else from plugins side, except what I already did...jmfowler wrote: Jimstar and TBFX - using 3.03f - good news and bad - I have studied the frames from last night, the first frame rendered in an hour, but the four frames after that took 1.5 hours meaning the 580 card ( probably this one as i have seen this one fail) failed on the second frame and the others, BUT the render kept going!!. NOW - I also had another render command in the .bat file and the render started on the same scene but at a different place - Again the first frame rendered in 1 hour but the frames after that took 1.5 hours...
So - jimstar's workaround fix appears to keep the render going with a failed card but I still don't know why my 580 card fails after the first frame???.....when I render the same scene on my other
computer using 3*680 cards it takes an hour for the image - so I am positive the 580 works all the way through on the first frame, plus the 1.5 hours versus 1 hour is another sign.
I'm going to make my .bat file with a command line for each frame to see if I can keep all 3 cards going.
Jimstar any thoughts?
I think that all is OK in the next render command due to full reloading of engine module, so during initializing the engine reinitializes all cards (it does not do it when the plugin is going to render just the next frame). So, if you will render every frame in separate render command - the render will last much longer, as the engine module will be reloaded every frame...
Yes, thank you, there was the bug. Will be OK in next version.TBFX wrote:I was doing some virtual shader tests on a simple particle instance scene and have come across some weird behavior. It seems depending on the method of cycling through different geometry on the instances Octane is rendering different geo to what is shown in the viewport. I've attached the scene.
If you go to the instancing section of the particle shape attribute editor and change the cycle options age attr from Birth Time to Particle ID you will get a different result but one that still does not match the viewport.
JimStar wrote:It is the engine issue, I can't do anything else from plugins side, except what I already did...jmfowler wrote: Jimstar and TBFX - using 3.03f - good news and bad - I have studied the frames from last night, the first frame rendered in an hour, but the four frames after that took 1.5 hours meaning the 580 card ( probably this one as i have seen this one fail) failed on the second frame and the others, BUT the render kept going!!. NOW - I also had another render command in the .bat file and the render started on the same scene but at a different place - Again the first frame rendered in 1 hour but the frames after that took 1.5 hours...
So - jimstar's workaround fix appears to keep the render going with a failed card but I still don't know why my 580 card fails after the first frame???.....when I render the same scene on my other
computer using 3*680 cards it takes an hour for the image - so I am positive the 580 works all the way through on the first frame, plus the 1.5 hours versus 1 hour is another sign.
I'm going to make my .bat file with a command line for each frame to see if I can keep all 3 cards going.
Jimstar any thoughts?
I think that all is OK in the next render command due to full reloading of engine module, so during initializing the engine reinitializes all cards (it does not do it when the plugin is going to render just the next frame). So, if you will render every frame in separate render command - the render will last much longer, as the engine module will be reloaded every frame...
Ok - thank you jimstar, your fix is at least keeping the render going and my render last night with separate command lines for each frame was pretty good - because my frames take about an hour each (for 3 cards) the reloading isn't too much of a big deal at the moment, do you know if the octane team is aware of this issue? - I'm hoping this might be fixed soon.
Cheers,
JF.
I
Maya 2013 / W7 64bit / Intel i7 980 / 24GB / Gigabyte gtx 580 3GB + 2*EVGA gtx 680 4gb / MB Gigabyte X58A-UD3R / NZXT Phantom - white/black case
Maya 2013/W7 64bit/Intel i7 3820/16GB/ MB Gigabyte GA-X79-UD3 /4*EVGA gtx 680 4gb/SSD 6Gb.s /CM HAF X case
Maya 2013/W7 64bit/Intel i7 3820/16GB/ MB Gigabyte GA-X79-UD3 /4*EVGA gtx 680 4gb/SSD 6Gb.s /CM HAF X case
Both 2012/2013 installed into c:\program files\autodesk, and it's default path. Also i checked the registry keys and they're ok.
The problem is that if you have both maya versions installed, installer offers you to install octane plugin just for 2013, 2012 is unchecked by default.
I can turn it on manually, just thought it would be nice to do that automatically as it supposed to be.
PS: Jim, i sent you PM - i was able to render these scenes just fine with 2.58 version. So it seems it's not octane engine eats all the memory.
The problem is that if you have both maya versions installed, installer offers you to install octane plugin just for 2013, 2012 is unchecked by default.
I can turn it on manually, just thought it would be nice to do that automatically as it supposed to be.
PS: Jim, i sent you PM - i was able to render these scenes just fine with 2.58 version. So it seems it's not octane engine eats all the memory.
Freelance CG TD, developer.
Now I've understood correctly. Just from your previous explanation "neither of the latest installers was able to detect maya 2012 installation, when 2013 is installed along with 2012" I understood that it just does not install the 2012 to the right place when you choose the 2012 checkbox.Joss wrote:Both 2012/2013 installed into c:\program files\autodesk, and it's default path. Also i checked the registry keys and they're ok.
The problem is that if you have both maya versions installed, installer offers you to install octane plugin just for 2013, 2012 is unchecked by default.
I can turn it on manually, just thought it would be nice to do that automatically as it supposed to be.
It was made such way by intention, just most recent version is choosen by default... I will think about changing it...
I already answered you: It is Octane engine problem. And I already wrote you that all the memory which this plugin uses for mesh data during compilation of your scene is just about 60MB.Joss wrote:PS: Jim, i sent you PM - i was able to render these scenes just fine with 2.58 version. So it seems it's not octane engine eats all the memory.
2.58 plugin uses another engine SDK (2.58 engine). This version uses 3.03 SDK. The huge memory consuming starts inside the voxelizing SDK function after it is called by plugin, and I'm forced to kill the Maya's process before it is returned from this function due to all memory is out and all gets stuck.