The main parameters of the three methods are compared as follows:
650) this.width=650; "Title=" vgpu2. JPG "src=" http://s1.51cto.com/wyfs02/M00/78/B0/wKioL1aBRMugejAwAAI30P2uK8A079.jpg "alt=" Wkiol1abrmugejawaai30p2uk8a079.jpg "/>
Three ways to support the model list of GPUs :
650) this.width=650; "Title=" VGPU3. JPG "src=" http://s1.51cto.com/wyfs02/M02/78/B0/wKioL1aBRV3BRB0gAAF6W6NvrhI673.jpg "alt=" Wkiol1abrv3brb0gaaf6w6nvrhi673.jpg "/>
VGPU different profile combinations in NVIDIA K1and K2 :
650) this.width=650; "Title=" Vgpu4. JPG "src=" http://s3.51cto.com/wyfs02/M01/78/B2/wKiom1aBRjmyCUQDAAI7-8keIG4933.jpg "alt=" Wkiom1abrjmycuqdaai7-8keig4933.jpg "/>
VGPU combination of methods:
VGPU at the virtual machine, the core of each GPU can only establish a VGPU configuration, for example: If a K2 core establishes a virtual machine using the k240q configuration file, then this core will no longer be able to establish a virtual machine configuration different from k240q. This means that the core can only build 3 k240q configured VMS, and cannot build k260q,k280q or k220q configured virtual machines.
The following is an image of the NVIDIA K2 GPU configuration file:
650) this.width=650; "Title=" Vgpu5. JPG "src=" http://s3.51cto.com/wyfs02/M02/78/B0/wKioL1aBRiPhZdieAAHHkAdic9I892.jpg "alt=" Wkiol1abriphzdieaahhkadic9i892.jpg "/>
Note: The red allocation method is not allowed.
The next section describes the commonly used link addresses in the VMware VGPU Scenario
This article is from the "Chenzhonghua" blog, make sure to keep this source http://chenzhonghua.blog.51cto.com/128732/1729345
VMware GPU Virtualization Technical parameters