NVIDIA
using grid cards in a render farm
can GRID be used in a non virtual machine environment thanks Jim
can GRID be used in a non virtual machine environment

thanks
Jim

#1
Posted 04/26/2016 02:24 PM   
Yes it can. The GRID SDK allows a range of technologies to integrate. Quadro products may also be an option for you to consider.
Yes it can. The GRID SDK allows a range of technologies to integrate. Quadro products may also be an option for you to consider.

#2
Posted 04/26/2016 06:01 PM   
Hi Rachel, thanks for the quick response. i admin a small 3dsmax render farm for my company consisting of a master machine and approx 60 slave machines over a corporate network. my IT manager has offered 4No K2 grid cards to me (that were for another project that was mothballed) in return for removing a large number of the existing slave machines. he does not want to run virtual machines, but thinks that the grid cards could be put into slave machines and use the cuda cores to do the rendering. i have looked everywhere, but have been unable to find anyone that has done anything similar. do you think this is possible or is my IT manager way of track? Thanks Jim
Hi Rachel,
thanks for the quick response.
i admin a small 3dsmax render farm for my company consisting of a master machine and approx 60 slave machines over a corporate network. my IT manager has offered 4No K2 grid cards to me (that were for another project that was mothballed) in return for removing a large number of the existing slave machines. he does not want to run virtual machines, but thinks that the grid cards could be put into slave machines and use the cuda cores to do the rendering.
i have looked everywhere, but have been unable to find anyone that has done anything similar.
do you think this is possible or is my IT manager way of track?

Thanks
Jim

#3
Posted 05/03/2016 02:40 PM   
the K2 cards are very specialist server cards each with 2xk5000 like spec'd GPUs, so I'd check the numbers on sheer cores your manager wants to replace. The first step would be to check the machines you want to put the cards in are certified: http://www.nvidia.com/object/grid-certified-servers.html After that if you want to avoid a hypervisor you are talking physical and each GRID card has multiple GPUs (on a hypervisor you can have multiple Vms so one to each GPU). I'm not sure what the options for multiple GPUs on physical and then presumably someway of remoting/uploading the rendering jobs are myself but I'll ask my colleagues who do more with physical. Since rendering is often CUDA/OpenCL on a hypervisor this would have to be pass-through. But I think you are askign about physical anyway. Rachel
the K2 cards are very specialist server cards each with 2xk5000 like spec'd GPUs, so I'd check the numbers on sheer cores your manager wants to replace.

The first step would be to check the machines you want to put the cards in are certified: http://www.nvidia.com/object/grid-certified-servers.html


After that if you want to avoid a hypervisor you are talking physical and each GRID card has multiple GPUs (on a hypervisor you can have multiple Vms so one to each GPU). I'm not sure what the options for multiple GPUs on physical and then presumably someway of remoting/uploading the rendering jobs are myself but I'll ask my colleagues who do more with physical.

Since rendering is often CUDA/OpenCL on a hypervisor this would have to be pass-through. But I think you are askign about physical anyway.
Rachel

#4
Posted 05/04/2016 09:14 AM   
Hi Jim, what type of renderer are you planning to work with please. Once we know that we can hopefully point you in the right direction.
Hi Jim, what type of renderer are you planning to work with please. Once we know that we can hopefully point you in the right direction.

#5
Posted 05/04/2016 10:21 AM   
Hi All, the Render Farm uses Autodesk Backburner to issue jobs to the farm. the plan at the moment is to put the K2 cards into workstations (Dell T3600 and T3500), one per workstation. The manager machine for Backburner is a small low power machine that is only used as a controller, this could be used for the Hypervisor or other software to control the systems if required. the Renderer is currently mental Ray, but for this setup we would be using Vray. cost savings on reduced machines would allow us to purchase this. What i have laid out above is open to some modification, but funds are tight and we are trying to make do with what we have. The reason for trying to downsize the farm from 60 slave machines to 4 GRID machines is cost. Each machine currently costs about a grand a year to be on the network plus electric and added pressure on the air con(they do produce quite a bit of heat on full render mode), when you have staff using these machines during the day this cost is negated. Thanks Jim
Hi All,

the Render Farm uses Autodesk Backburner to issue jobs to the farm.
the plan at the moment is to put the K2 cards into workstations (Dell T3600 and T3500), one per workstation.
The manager machine for Backburner is a small low power machine that is only used as a controller, this could be used for the Hypervisor or other software to control the systems if required.

the Renderer is currently mental Ray, but for this setup we would be using Vray. cost savings on reduced machines would allow us to purchase this.

What i have laid out above is open to some modification, but funds are tight and we are trying to make do with what we have.
The reason for trying to downsize the farm from 60 slave machines to 4 GRID machines is cost.
Each machine currently costs about a grand a year to be on the network plus electric and added pressure on the air con(they do produce quite a bit of heat on full render mode), when you have staff using these machines during the day this cost is negated.

Thanks
Jim

#6
Posted 05/09/2016 09:10 AM   
Hi Jim, so long as you are talking about VRay-RT (which is GPU accelerated) version of the Chaos product then what you want to do should be feasible (Vray itself is a pure CPU renderer). As Rachel suggests if the K2's are installed in suitable hardware with windows (or for that matter even Linux) installed as the OS then the GPU's on the K2 will appear as regular GPU's in device manager and should be visible to VRay-RT. Note that we don't QA or certify this configuration for K2 so doing some testing before full purchase/deployment is advised.
Hi Jim, so long as you are talking about VRay-RT (which is GPU accelerated) version of the Chaos product then what you want to do should be feasible (Vray itself is a pure CPU renderer). As Rachel suggests if the K2's are installed in suitable hardware with windows (or for that matter even Linux) installed as the OS then the GPU's on the K2 will appear as regular GPU's in device manager and should be visible to VRay-RT. Note that we don't QA or certify this configuration for K2 so doing some testing before full purchase/deployment is advised.

#7
Posted 05/10/2016 10:54 AM   
Hi, thanks for the info. looks like i have some testing to do. i will report back when i have some answers or more problems. cheers Jim
Hi,
thanks for the info.
looks like i have some testing to do.
i will report back when i have some answers or more problems.
cheers
Jim

#8
Posted 05/19/2016 08:03 AM   
In a recent webinar a similar enquiry came up... Q: Is GRID is a viable solution for my Creative Team whom are using mostly software like Adobe Aftereffects, 3DS MAX, V-ray etc. ? NVIDIA’s GRID solution is applicable to any situation where applications running in a virtual environment require graphical acceleration to perform well. This statement applies across the range of applications typically found in M&E (Media and Entertainment) environments like 3dsMax, Maya, Premier Pro (PPro), Aftereffects, Photoshop, Nuke etc., with one or two provisos. The most notable proviso is where CUDA or OpenCL support is required by an application. Since CUDA is only supported today by our 8Q profiles (where the whole physical GPU is allocated to the VM) applications that require CUDA will need to use this profile with the attendant reduction in density possibilities. Applications that do use CUDA in an M&E context are PPro (the Mercury playback engine), Aftereffects (3D text rendering) and many rendering engines like VRay-RT. So these applications either need to be used in such a way that the CUDA part of the application is not used (for example not everyone needs 3D text in Aftereffects) or the CUDA part of the app is allowed to fall back to CPU. For example the Mercury Playback engine can use CPU and that would likely be acceptable for situations where heavy effects or multiple video streams where not being processed. Other considerations for M&E would be that Wacom tablet support is only available in a Citrix environment and is not fully certified. Additionally the color-space support in VDI protocols is often not sufficient for true color QA or colour grading applications. The remoting protocols and virtualisation stacks also handle audio/graphical channel synchronisation differently and users should consult their remoting vendor e.g. Citrix/VMware/other on their technologies to avoid drift over time. Many M&E applications run under Linux and this is a popular choice for many M&E customers. It should be noted that full Linux support with vGPU is available with both VMWare and Citrix (introduced in XenServer 7.0 just after the webinar was broadcast). RGS, XenDesktop, Horizon, MechDyne TGX and NiceDCV all support vGPU enabled VM’s in their protocols too. Which Linux OSs are supported varies across vendors so those evaluating should investigate each protocols support. A good range of M&E requirements can be met by GRID technology with some of our customers successfully using Maya, 3DSMax and similar but some aspects of a workflow may still require a dedicated workstation, colour grading complex editing with PPro or final QA for example. One of our rendering and virtualization experts is having a similar discussion with a similar customer on this thread https://gridforums.nvidia.com/default/topic/801/using-grid-cards-in-a-render-farm/ and if you want to ask more questions that thread could be an appropriate place to do so.
In a recent webinar a similar enquiry came up...

Q: Is GRID is a viable solution for my Creative Team whom are using mostly software like Adobe Aftereffects, 3DS MAX, V-ray etc. ?

NVIDIA’s GRID solution is applicable to any situation where applications running in a virtual environment require graphical acceleration to perform well. This statement applies across the range of applications typically found in M&E (Media and Entertainment) environments like 3dsMax, Maya, Premier Pro (PPro), Aftereffects, Photoshop, Nuke etc., with one or two provisos.

The most notable proviso is where CUDA or OpenCL support is required by an application. Since CUDA is only supported today by our 8Q profiles (where the whole physical GPU is allocated to the VM) applications that require CUDA will need to use this profile with the attendant reduction in density possibilities. Applications that do use CUDA in an M&E context are PPro (the Mercury playback engine), Aftereffects (3D text rendering) and many rendering engines like VRay-RT.

So these applications either need to be used in such a way that the CUDA part of the application is not used (for example not everyone needs 3D text in Aftereffects) or the CUDA part of the app is allowed to fall back to CPU. For example the Mercury Playback engine can use CPU and that would likely be acceptable for situations where heavy effects or multiple video streams where not being processed.

Other considerations for M&E would be that Wacom tablet support is only available in a Citrix environment and is not fully certified. Additionally the color-space support in VDI protocols is often not sufficient for true color QA or colour grading applications. The remoting protocols and virtualisation stacks also handle audio/graphical channel synchronisation differently and users should consult their remoting vendor e.g. Citrix/VMware/other on their technologies to avoid drift over time.

Many M&E applications run under Linux and this is a popular choice for many M&E customers. It should be noted that full Linux support with vGPU is available with both VMWare and Citrix (introduced in XenServer 7.0 just after the webinar was broadcast). RGS, XenDesktop, Horizon, MechDyne TGX and NiceDCV all support vGPU enabled VM’s in their protocols too. Which Linux OSs are supported varies across vendors so those evaluating should investigate each protocols support.

A good range of M&E requirements can be met by GRID technology with some of our customers successfully using Maya, 3DSMax and similar but some aspects of a workflow may still require a dedicated workstation, colour grading complex editing with PPro or final QA for example.
One of our rendering and virtualization experts is having a similar discussion with a similar customer on this thread https://gridforums.nvidia.com/default/topic/801/using-grid-cards-in-a-render-farm/ and if you want to ask more questions that thread could be an appropriate place to do so.

#9
Posted 06/21/2016 10:18 AM   
Scroll To Top

Add Reply