Environment Details:
Cisco USCS-C240-M3S
VMware ESXi, 6.0.0, 3620759
GRID K1 - Driver version 367.92
Guest Driver: 369.95
Horizon View 7.1.0
CADD Based VM Build:
Windows 7
4vCPU
12GB RAM
grid_K160Q Profile
Client Version: 4.4.0 build-5171611
Hello,
I am looking for some best practices on setting my Group Policies for the Blast Extreme and PCoIP Session Settings. I have handful of vGPU VMs that are running 3d CAD workloads. For the most part, they seems to working out failly well for the users. I have users both internal and external users accessing them. No major bandwidth constraints at the Datacenter where these are being hosted. However the end users accessing them maybe on varying connecting speeds.
Also, I decided to run a FurMark test on one of them and notice it was only getting around 9FPS. I was expecting something a bit higher than this. I was also monitoring the GPU inside of the VM session using GPU-Z and noticed that the gpu load did not move past 0%. However, while monitoring the host using nvidia-smi I observed one of the cores on the K1 card went to 99%. I am not sure if GPU-Z is properly detecting the furmark workload during the test or not. Also, while monitoring the the gpu load via GPU-Z while in a CAD application, the GPU-load does not budge either. If their are other tools to monitor the GPU load inside the guest VM I would be excited to hear about them.
Thank you for any additional details anyone can provide.
Cheers
