OTOY Forums

hi

I often work from home, and I wonder if I can use my computer at work as a render slave?
I can connect via vpn and Remote Desktop.

Has anyone experience with this? Does it require an additional license?

Thanks

I do not have experience with VPN, just using home network.
But you will need additional Standalone licence for every slave.

thanks rappet

ok, i hope that someone can confirm that it will work, before i buy an additional license.

I regularly work with octane through a remote connection, although I'm working fully remotely; not trying to use my remote computers as extra render nodes. If you haven't tried working remotely with Octane yet, one thing to note is that if you are using Windows RDP, it disables all CUDA functionality. Windows RDP loads its own set of graphics drivers specifically optimized for RDP, which is why it works so much faster than other option (there are work arounds to this though).

It took me a while to realize that, so I just thought I'd point that out and hopefully save you some headaches if you were trying to use RDP

hi itsallgood9

thanks for the information - i was aware of that - tried to render via remote desktop, but found the same issue as you.

so what i would like to do, is to render on my pc at home, and use the two 780 ti cards at work, to speed up rendertime - if possible

thanks

I'd say technically this is possible, but practically it's not going to work.
The problem will be the bandwidth between your home and work machine.

Obviously, network rendering works fine on a gigabit ethernet LAN.
You can also get the slave to connect to a master by specifying the IP address and port of the master (so you don't actually have to use the slave daemon or be on the same LAN). So long as there is a traffic route to/from the master the slave should connect.

The problem will be transferring the scene data over the internet to your work machine and getting the results back.
Everything that ends up in the VRAM of your local GPU will end up needing to be sent over the network for the slave to start working on it. I believe there are some optimizations but in general if you have a 4GB scene then most likely you will need to transfer upwards of 4GB to the slave.

So unless you have a lightning fast network/internet connection to your work machine, I don't think it will really work out...

Ah ok.

Thanks for the explanation fooze.

Ise remote accessing software (VNC like) & setup something like DropBox - You can tweak the scene, (use all heavy materials & models already synchronised - placed on dropbox folder) & then only upload pretty light .OCS file (made on standalone or exported from Your app..) In this case You can fire up remote desktop once scene is done, fire the scene to render & keep working on other shot with Your computer. You basically split resources for two works. Not so elegant solution, but it might work bit better than using "network rendering".

Hi glimpse

Thanks for your tip.

I hoped I could use the two cards at work in combination the my card, so I had three cards instead of only one.

But it seems that it get more complicated than I hoped.

glimpse wrote:Ise remote accessing software (VNC like) & setup something like DropBox - You can tweak the scene, (use all heavy materials & models already synchronised - placed on dropbox folder) & then only upload pretty light .OCS file (made on standalone or exported from Your app..) In this case You can fire up remote desktop once scene is done, fire the scene to render & keep working on other shot with Your computer. You basically split resources for two works. Not so elegant solution, but it might work bit better than using "network rendering".

Do you know if I could setup this same thing on an Amazon EC2 remote desktop, not the Octane AMI but just a regular EC2. Then, run multiple instances? Wish I knew why I can't find any information about this. I know this is what the Octane AMI is eventually supposed to do but I can't figure out what is stopping us all from doing it ourselves right now.