Improve vGPU Selection

[onenhansen]

Description Allow option to select vGPUs on different devices when multiple are assigned to one VM

Enhance scheduler functionality:

• allow placement rules for vGPU's, packing/striping for systems with multiple GPUs
• increase awareness of GPU capacities and usage with currently deployed profiles.
• discover homogeneous vs heterogenous mode on the GPU ( homogenous mode: the first profile allocated determines all profiles for that card until the last one is removed )
• Allow PROFILE selection along with SHORT_ADDRESS

Use case When vGPU's are time-sliced they are competing in time for each other rather than being split across multiple devices. Systems with multiple GPU's may not properly allocate vGPU's from the second card depending on the profile(s) used on the first card

Interface Changes Perhaps a checkbox to turn this feature on/off in the VM template editing page.

Progress Status

• [ ] Code committed
• [ ] Testing - QA
• [ ] Documentation (Release notes - resolved issues, compatibility, known issues)