
Alternative method for animations?
Forum rules
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
Please add your OS and Hardware Configuration in your signature, it makes it easier for us to help you analyze problems. Example: Win 7 64 | Geforce GTX680 | i7 3770 | 16GB
- Jaberwocky
- Posts: 976
- Joined: Tue Sep 07, 2010 3:03 pm
How about the exporter , exporting all the frame parameters from the application in 1 hit in one large obj file.Then a small application called an Octane server running on the computers CPU could schedule to feed each frame straight to the GPU on the card.The finished frame is sent back to the Server on the CPU,its then renderered out to an image on the hard disk and then the next one is sent to the GPU.All You would then need to do is set up the process,shut down the 3D application,get the server started,watch to make sure the process is running and then go away and let it run through the frames.Two tightly integrated & coded apps,Octane on the GPU and octane Server running on the CPU should certainly speed things up a lot. 

CPU:-AMD 1055T 6 core, Motherboard:-Gigabyte 990FXA-UD3 AM3+, Gigabyte GTX 460-1GB, RAM:-8GB Kingston hyper X Genesis DDR3 1600Mhz D/Ch, Hard Disk:-500GB samsung F3 , OS:-Win7 64bit
to know the export time with the standard build you can use the script (somewhere i found also a good workin 2.5 version)face wrote:I haven´t the intention to build it.mib2berlin wrote:Special Blender build:*
V-Ray/Blender 2.49.12 (Windows 32bit)
V-Ray/Blender 2.49.12 (Linux 64bit)
Source code
* Special Blender 2.49b build with mesh export optimization (up to 10 times faster export).
Cheers mib
I don´t know the export time with the standard build.
face
if you have any win 32bit os available you can also try the optimized 32bit binaries.
if there is a car using some different technic, why not try driving it and maybe look into the motor?
we would like to see the export speed up, so maybe also python is one bottleneck..
- Mint 10 64bit nvidia drv 260.19.29/cudatoolkit3.0 intel q6600, 4gbRAM, GTX470 1,2GB
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
Thats correct, i agree with you.dave62 wrote:to know the export time with the standard build you can use the script (somewhere i found also a good workin 2.5 version)
if you have any win 32bit os available you can also try the optimized 32bit binaries.
if there is a car using some different technic, why not try driving it and maybe look into the motor?
we would like to see the export speed up, so maybe also python is one bottleneck..
And yes, i can install it, but you as a user which has use it, can say me maybe some times to export a 1mio tri polygon mesh.
I have looked into the source and haven´t seen any special workaround.
The python script uses the buildin export, right?
face
Win10 Pro, Driver 378.78, Softimage 2015SP2 & Octane 3.05 RC1,
64GB Ram, i7-6950X, GTX1080TI 11GB
http://vimeo.com/user2509578
64GB Ram, i7-6950X, GTX1080TI 11GB
http://vimeo.com/user2509578
hi,face wrote:Thats correct, i agree with you.dave62 wrote:to know the export time with the standard build you can use the script (somewhere i found also a good workin 2.5 version)
if you have any win 32bit os available you can also try the optimized 32bit binaries.
if there is a car using some different technic, why not try driving it and maybe look into the motor?
we would like to see the export speed up, so maybe also python is one bottleneck..
And yes, i can install it, but you as a user which has use it, can say me maybe some times to export a 1mio tri polygon mesh.
I have looked into the source and haven´t seen any special workaround.
The python script uses the buildin export, right?
face
if i find some sparetime i will do some tests concerning the speed dif
- Mint 10 64bit nvidia drv 260.19.29/cudatoolkit3.0 intel q6600, 4gbRAM, GTX470 1,2GB
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
An alternative way, for octane developers, is to implement just a simple text parameter import for each node.
each node could take for example an ascii text file input with a line per frame of node's parameter values.
i think this is much more easy for them to implement plus much more open for us to experiment with exporters etc.
each node could take for example an ascii text file input with a line per frame of node's parameter values.
i think this is much more easy for them to implement plus much more open for us to experiment with exporters etc.
scene 1.2mio faces
vrayexport otimized:19s
normal non optimized 2.49 build:800s
vrayexport otimized:19s
normal non optimized 2.49 build:800s
- Mint 10 64bit nvidia drv 260.19.29/cudatoolkit3.0 intel q6600, 4gbRAM, GTX470 1,2GB
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
Thanks, now triangulated, pleasedave62 wrote:scene 1.2mio faces

Time is enough, it´s lesser work

face
Win10 Pro, Driver 378.78, Softimage 2015SP2 & Octane 3.05 RC1,
64GB Ram, i7-6950X, GTX1080TI 11GB
http://vimeo.com/user2509578
64GB Ram, i7-6950X, GTX1080TI 11GB
http://vimeo.com/user2509578
the factor is about 40x times faster between 2.49 non opt and 2.5 opt.
maybe i should do another test. comparing only 2.5x versions. opt and non opt for vray.
the difference seem to be a lil to big imo atm, but maybe its just because its hardcoded wo python.
next step after this, comparison with obj/mtl export from blender.
maybe just some hardcoded obj export in blender would help a lot speeding up th whole render workflow.
imo i would even pay for the exporter to, if i got some great performance increase.
edit: i think the huge difference must depend on some other stuff..
here is the octane blender 2.5 export: 54s but at least half as fast as the optimized one for vray
maybe i should do another test. comparing only 2.5x versions. opt and non opt for vray.
the difference seem to be a lil to big imo atm, but maybe its just because its hardcoded wo python.
next step after this, comparison with obj/mtl export from blender.
maybe just some hardcoded obj export in blender would help a lot speeding up th whole render workflow.
imo i would even pay for the exporter to, if i got some great performance increase.
edit: i think the huge difference must depend on some other stuff..
here is the octane blender 2.5 export: 54s but at least half as fast as the optimized one for vray
- Mint 10 64bit nvidia drv 260.19.29/cudatoolkit3.0 intel q6600, 4gbRAM, GTX470 1,2GB
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
@geop
hqnormal where switched off in this "benchmark"
hqnormal where switched off in this "benchmark"
- Mint 10 64bit nvidia drv 260.19.29/cudatoolkit3.0 intel q6600, 4gbRAM, GTX470 1,2GB
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x
- Mint 10 64bit/ Win7 64bit nvidia drv 260.19.29/cudatoolkit3.2 amd X6, 16gbRAM, 2x GTX580 3GB
->Octane 2.44/ Blender2.5x