Information

1. Windows 7 Vmx File
2. Download Vmx File
3. How To Run Vmx File
4. Vmx Download Free

-Integration with Unidesk
-Supported Versions
-Licensing Considerations
-Prerequisites
-Manual Update
-VMX Update Utility
-Desktop Processing Utility
-NVIDIA Driver Layer
Overview

View and Download Asus M2N8-VMX user manual online. M2N8-VMX Motherboard pdf manual download. When using PCI cards on shared slots, ensure that the drivers support “Share IRQ” or that the cards do not need IRQ assignments; otherwise, conflicts will arise between the two PCI groups, making the system unstable and the card inoperable.

With the release of vSphere 6 VMware and NVIDIA have included a new feature set for shared GPU on Horizon View Desktops called vGPU. The advantages of vGPU compared with vSGA and vDGA can best be explained by understanding each technology.

Soft 3D

VMware View includes a software based GPU. When using the Soft 3D VMware allocates 128 MB of RAM for the video driver and uses the VM’s CPU to provide virtual GPU. The processor load is significant when for instance watching a video. This driver supports only up to DirectX 9 and OpenGL2.1. This is appropriate for non demanding graphical applications.

1. VXM Step Motor Controller. Velmex offers several VXM Stepper Motor Controllers to govern the performance of the motors available for Velmex products. They make precise positioning of the slides, stages and tables simpler and more effective.
2. Downloads Free! 43 Drivers, Utilities, Manual and BIOS for Asus P5GD1-VM Motherboards. Here's where you can download Free! The newest software for your P5GD1-VM.

vSGA
This provides the ability to share NVIDIA GRID processors among many virtual desktops. An NVIDIA driver is installed on the Hypervisor and the desktops use a proprietary VMware developed driver that will access the shared GPU. This option supports only up to DirectX 9 and OpenGL2.1
Requires	Benefits	Drawbacks
vSphere 5.1 U1 or later 5.5 recommended View 5.2 or later 5.3 recommended Max 512 MB RAM (half for Desktop Driver) Windows 7 or later	Up to 16 Concurrent Users per K2 card Up to 32 concurrent users per K1 card Up to 8 cards per host	Limited support for Directx and OpenGL

vDGA
This is a hardware pass-through mode where the GPU is not shared but accessed directly by the virtual desktop. This mode supports the real NVIDIA graphics driver and attaches directly to the GPU from the desktop. This option is very expensive as a GRID K2 card can only support 4 desktops and a K1 card only 2 desktops. This option supports DirectX 9,10, and 11, OpenGL 4.4 and NVIDIA CUDA 6.5 and should have the graphical performance of a high end graphical workstation.
Requires	Benefits	Drawbacks
Dedicated GPU vSphere 5.1 U1 or later 5.5 recommended View 5.2 or later 5.3 recommended Max 512 MB RAM (half for Desktop Driver) Windows 7x64 or later	Best in class Performance	Expensive Each GPU dedicated to a desktop

vGPU
vGPU has many of the benefits of vDGA but can also share the NVIDIA GRID GPU’s. Here you install an NVIDIA VIB driver in the hypervisor and an NVIDIA driver on the virtual desktop. vGPU supports DirectX 9,10 and 11 and OpenGL 4.4
Requires	Benefits	Drawbacks
vSphere 6.0 or later vCenter 6.0 or later Horizon View Agent 6.1, 6.2 or later recommended Hardware Version 11 or later • Full Memory Reservation	Up to 16 Concurrent Users per K2 card Up to 32 concurrent users per K1 card Up to 8 cards per host Drawbacks All Desktop Memory must be reserved DRS is not supported Snapshots are not supported so Unidesk backups are not supported

Back to top

NVIDIA vGPU Profiles

When using vGPU NVIDA support four different profiles. Each profile assigns more or less resources per user from a GPU. All the users of a single GPU on an NVIDIA GRID card must use the same profile. The profiles are listed below for both K1 and K2 cards. Note the Max Users per board.

Integration with Unidesk

It is fairly easy to integrate the NVIDA technology with Unidesk desktops. The NVIDIA driver can be installed into an application layer and therefore delivered to just a subset of your desktops.
The hardest part of the integration is that the shared pci device and several vmx settings must be added to each desktop for it to work.
For persistent desktops this can be achieved in several ways:

• Manually add the shared PCI device using the vCenter Web Interface
• Use the Unidesk provide mass edit of VMX files after the desktops are deployed but before they are used
• Integrate the Unidesk Desktop Processing Script that modifies desktop settings during the desktop build process.

For Non-persistent desktops there are only two ways to make these changes:
1. Disable the change to Non-persistent during the build, make the changes to the desktop using number 1 or 2 above, then delete the BlockReadyMessage registry key to finalize the Non-persistent Build.
2. Integrate the Unidesk Desktop Processing Script that modifies desktop settings during the desktop build process.
The Desktop Processing utility also allows you to configure other VM settings like:

• devices.hotplug = false which will disable the ability to remove the NIC or SCSI devices on a desktop by the user.
• CPU Reservations • Memory Reservations
• Put Desktop into q specific folder
• Put desktop into a specific Resource Pool

These options will be explained later in the document.

Supported Versions

This document has been developed and tested using vSphere 6.0, vCenter 6.0 and an NVIDIA GRID K1 card. I have seen Unidesk customers that have deployed with up to 4 K1 cards per host.

Licensing Considerations

vGPU requires vSphere Enterprise Plus and View Horizon 6.1 or later.

Layering Naming and versioning

Unidesk recommends including the OS Type and OS bit level in the name if you have multiple version of an OS, for Example NVIDIA Drivers Win7X64. In 2.0 the bit level is not critical because we will not allow assignment to the wrong OS layer but it still provides important information. For versions remember that when choosing a layer you can see the layer name and version name but not the version description. Use naming that will allow you to differentiate versions appropriately. For example while still in development/testing “1.0 12-12-14-2012 QA ONLY”, but when ready for production “1.0 12-12-2012”.

Prerequisites

You must decide before starting what method will be used to modify the VMX settings for desktops using vGPU. As stated earlier there are several choices for Persistent Desktops. These include in order of complexity manual update, the VMX Update Utility and the Desktop Processing Utility.
Back to top

Manual Update

If you will only be deploying a small number of persistent desktops with vGPU the easiest approach is to manually edit the desktop in the vCenter Web Client adding the shared PCI device. This is very easy to do:

1. Open the client
2. Navigate to the desktop
3. Shut Down the desktop
4. Click on Manage
5. Click on Edit Settings
6. In New Device Select Shared PCI Device and click Add Once you add the device you will see Select NVIDIA GRID vGPU and the desired GRID profile.
   Once you add the device you will see
   
7. Click Ok
8. Start The Desktop

Back to top

VMX Update Utility

If you will have many persistent desktops configured with vGPU you can use either of two utilities we have developed to modify VM settings on virtual desktops. The simpler approach is to use the Edit CMX Files Utility.
This utility will allow you to select any number of pre-created desktops and modify their VMX settings. To use the utility you first shutdown the desktop and put them in maintenance mode. Then you run the utility and modify their VMX settings.
To set up the settings in the utility after it is installed, open the exe and add your vCenter credentials.
Then in the Key Value area add the following:
VMXKey->MemReservationMB VMX Key Value->The memory size of your Desktops in MB
Click Add VMX
Value VMXKey-> pciPassthrough0.virtualDev VMX Key Value-> vmiop
Click Add VMX
Value VMXKey-> pciPassthrough0.vgpu VMX Key Value-> GRID Profile e.g. grid_k120q
Click Add VMX
Value VMXKey-> pciPassthrough0.present VMX Key Value-> TRUE
Click Add VMX Value
Once these settings are set you click GetVMs to catalog all the virtual desktops in vCenter. Then either click search with no wildcard pattern to see all VMs or enter a wildcard pattern and click search. This will add a list of VMs matching the search criteria. The text entered in the search box will be found anywhere in the VM name. Do not include wildcard character. To Find all machine that include UDM for example just enter UDM and click search.
Then select the desired machines and click save. If you are adding to an existing list check append and click save. The desktops will then show up in the Action Area. Make sure the list is correct and click Update VMX.
The script will make the appropriate changes to the VMX files.
See the included documentation for details on installing and using the utility.
Back to top

Desktop Processing Utility

The desktop processing utility is more complex to setup and use but it provides the ability to modify the desktops configuration automatically during the build process. It also provides the ability to move a desktop into a particular folder or resource pool or both. The utility settings are based on configuration rules defined by desktop naming convention so you can apply different settings to different desktops based on naming standards.
The basic design of this utility is that during the desktop build the desktop a flag file with the desktop name and domain is written to a file share and the desktop is shut down. A script is run either in a loop or in a scheduled tasks that checks the share for flag files and if it finds one it makes the modification to the desktops VMX file and start the desktop back up so the build can complete.
In rode to use this utility the View Pool must be configured to “Take no Power Actions” and there must be at least one unassigned desktop in each View Dedicated Pool and one Available desktop in any View Floating pool. This is because if there is not View will automatically start one desktop in the pool which will break the script from working on the first desktop created. This utility is set up and configured in pretty much the same way as the VMX Update Utility though it’s a bit easier because all of the vGPU settings are already built in.
Here you enter a wildcard naming for example for desktops starting with NVk120-
NVK120-*

Windows 7 Vmx File

NVK180-*
To always set the same settings just use an * (asterisk) for the wildcard setting.
Then add the appropriate VMX settings for that configuration. Again the required settings are:
Use appropriate memory reservation and NVIDIA profile.
Then there is a new Unidesk Optimization utility and KMSSETUP.cmd that must be used to provide the build scripting in support of this utility. As of this writing this functionality is not in the distributed version of the Unidesk Gold Tools but it will be added at a later date. It is easy to tell if your version of the optimizer and kmssetup.cmd support this. If the optimizer has a check box in the top section that includes Desktop Processing
If the KMSSETUP.CMD in the C:windowssetupscriptskmsdir folder contains a section for
If you don’t have these they are downloadable with the utility from the Desktop Processing Utility.
See the documentation for details on installing this utility but the new optimizer and kmssetup.cmd should be added to a new version of the OS layer and then when creating the NVIDIA Driver layer the optimizer is used to create the CreateFlag.cmd and a flag file called Add_Restart that tell kmssetup to add an extra restart before the removeready is run on a Non Persistent desktop.
Back to top

Non persistent Desktop Considerations

As stated earlier for Non-persistent desktops the PCI hardware and GPU must be added by Windows before the desktop is converted from persistent to non-persistent or the desktop will ask for a reboot on every logon. There are two possible ways to support this. Use the optimizer and kmssetup discussed earlier and choose Disable RemoveReady for No desktops
This will stop the build process at the end leaving the NP desktops in a persistent state. You can then use any of the methods as discussed for persistent desktops and when complete navigate to the windowssetupscriptskmsdir folder and run the removeready.cmd which will delete the BlockReadyMessage registry entry and convert the desktop to nonpersistent. Be aware the desktop creation task will not complete until this happens.
Most customers will want to configure the Desktop Processing utility when using NP desktops. Just be aware that when doing so View Power Management cannot be used and Unidesk Power Management is required. This is because if View Power Management is enabled the desktops will not be able to be shut down during the build to update the VMX. However remember for NP desktops you must use one power management or the other or the desktops will never be restarted after a logout or shutdown.

VMware View Pool Pre-Requisite

Download Vmx File

If you will deploy some desktops with vGPU and others without it you will need separate View Pools to support both. The setting required to support vGPU is 3D Renderer : NVIDIA GRID VGPU
If using the Desktop Processing Utility you must also have Remote Machine Power Management : Take no power Action.

How To Run Vmx File

Back to top

Vmx Download Free

NVIDIA Driver Layer

The high level process for layering the vGPU is as follows:

1. Create View Layer
2. Create Install Machine with Shared PCI Device
3. Create NVIDIA Driver Layer
4. Test

Installation Steps For NVIDIA Driver Layer

1. Add Installation machine for NVIDIA only
   1. We will keep this as a separate IM for when we need to work on NIVIDIA
   2. Upgrade VM Compatibility to vSphere 6 (HW Version 11) if not already at this version
   3. Modify the install machine to add a “shared PCI Device” See pre-requisite section for details
   4. Choose desired NVIDIA Profile
2. Create NVIDIA Driver App Layer Using IM
   1. If the IM asks for a restart due to the HW upgrade then restart
   2. If Using the Desktop Processing Script Run the new optimizer select “Check to Use Desktop Processing…” Fill in the information. If you don’t want your users to see a “you must restart” message choose Additional Restart. Then use the save settings a-j at the end of the same section. Don’t use the save at the bottom of the optimizer as that will resave the optimizations.cmd which we only want in the OS layer. See Desktop Processing section above for details.
3. Copy NVIDIA driver locally to the IM (Its small)
4. Run Driver install “As administrator'
   1. You can use express or custom I always use custom to see what’s installing
      
5. Restart IM
6. Login. Once the driver install the mouse no longer works. You can use tabs to go into control panel and open the NVIDIA Control Panel. If it open the install was likely successful. If it wont open there is likely an issue.
7. Reboot
8. Finalize the layer
9. Create a View Pool as described
10. For Non Persistent Desktops Create a Unidesk Template and Enable Power Management.
11. Create a test desktops with just the OS, View Agent and NVIDIA layers
    1. Ensure memory matches what is defined in the desktop processing configuration for memory reservation
    2. Note View Power Management while set to Take no power action will still power on machines if there are no available desktops in a floating pool or no unassigned desktops in a dedicated pool. Therefore the first build will not run the DP script correctly because the desktop will not stay powered off. You must ensure that there is always at least one available powered on desktop in a view floating pool when creating desktops otherwise be sure to create more than 1 and remove the first added because it will not build correctly.
    3. You can only test with a View Client after login on right click on the desktop and run the NVIDIA Control Panel. Then click on Adjust Image settings with preview. If you see the rotating image you were successful
       

Updates

None of these drivers have automatic updates to deal with.