What version of Premiere (be specific like Premiere Version:, 18.104.22.168)?
We are on the CC version.
the GPU is NVidia 700 X serie.
we are running this on ESXI 6.0 VM Machine with Windows 10.
does the windows install see the nvidia card and nvidia drivers installed from the web?
do you have vmware set to allow the win10 vm to have direct access to the nvidia card?
if you have an intel cpu in the computer, make sure it supports vt-d and is enabled in the bios. amd equivalent is named amd-vi
VTD-I is enabled and I see passhthough device.
if I add the external VGA to the VM I will be able to use the " Mercury Playback Engine GPU Accelerator " ?
VMware is on and can face the internet no issue with facing the internet .
Do you maybe know how to specify which GPU is the VM going to use ?
for premiere to enable "mercury playback engine", it needs access to your gtx card with nvidia drivers installed. if you follow the kb article i linked above, it should walk you thru adding the gtx card to your vm. there is a good chance esxi won't allow this to work properly with your gtx 700 series card. you may have to look into using a quadro or possibly using another hypervisor that will allow passthrough with gtx cards. Passthrough Nvidia graphics card to VM | VMware Communities
you might want to ask on vmware's forums to ask how to get cuda program support and/or passthrough with a gtx, they should have better answers.
I managed to have the Passthought of the VGA however the VM doesn't recognize it
VGA is NGforce GT 218
any suggestions how to make it works ?
its gives Error Code 42
Guess what? The GT218 GPU (a/k/a the GeForce G210) is now too old to be supported in CC2015. And even if it were technically supported, most (though not all) cards that use that GPU had only 512 MB of graphics RAM - which is too little to even enable GPU acceleration at all. (Premiere Pro requires a minimum of at least 765 MB of free, unused graphics RAM to even enable MPE GPU acceleration at all - and that Windows itself consumes 20 MB of graphics RAM, which completely rules out 768MB cards.) Worse, the GT218 GPU has only 16 CUDA cores and an horribly low graphics memory throughput of no more than 12.8 GB/second (or, if your card has only DDR2 graphics memory, that throughput is an extremely abysmal 8 GB/second with only a 64-bit graphics memory bandwidth!). As such, its performance is so glacially slow that it will be slower with GPU acceleration enabled than a PC with a higher-end GPU that's running in software-only mode - and that the GT 218 will also substantially slow down performance in software-only mode as well!
Also, the older and newer GPUs conflict with one another. Tesla-architecture GPUs do not play nice with Kepler- or Maxwell-generation GPUs at all when asked to perform CUDA duties.
Thank you so much.
Can you suggest a GPU that will works fine on VMware ? and using the passthrough ?