嗨...
您有两种不同的选择。
您可以以直通模式（独立或作为XenApp安装）运行Windows服务器，在这种情况下，它需要驱动程序作为GPU（NVGRID系列），或者如果作为VM桌面运行，每个使用vGPU，它需要一个不同的驱动程序，
将与XenServer 6.2或6.5的下拉菜单部分（NVIDIA vGPU）中的XenServer捆绑在一起。
您需要单独安装VM驱动程序以及XenServer驱动程序。
GPU版本可以独立运行，不仅支持GRID，还支持许多不同的GPU，或者您可以在XenServer下将其作为VM运行，在这种情况下，您需要安装XenServer部分以便正确运行
将GPU直通分配给VM。
所有下载都在
http://www.nvidia.com/Download/index.aspx?lang=en-us。
需要哪些驱动程序的组合有点令人困惑。
至于偏好，一切都取决于。
如果你想要为每个桌面提供专用的，有保证的GPU数量，那么vGPU选项可能是更好的选择。
如果你希望任何人能够在不那么繁忙或更少的时候利用具有更多可用功率的GPU，那么GPU直通可能更适合。
有多个工程师可以分配，你甚至可以将两者结合起来。
至于AutoCAD，它的工作方式很好。
另请参阅http://blogs.citrix.com/2014/05/28/configuring-virtualised-autodesk-and-similar-applications-including-xenapp-gpu-sharing-中有关GPU直通AutoDesk应用程序的特定文章。
and-tools-for-sanity-checking-directxopengl-usage /？_ ga = 1.146046580.1558783495.1433165527（感谢那篇文章，Rachel！）。
您可以支持多少个实例将再次依赖于VGPU，如何分配分配，或者在GPU直通的情况下，依赖于不同用户可能同时执行的操作。
最好的祝福，
以上来自于谷歌翻译
以下为原文
Hi...
You have two different options. You can run the Windows server in passthrough mode (standalone or as a XenApp instalnce) in which case it will need a driver as a GPU (NVGRID Series) or if run as VM desktops using a vGPU for each, it needs a different driver that will be bundled with the XenServer in that dropdown menu section (NVIDIA vGPU) for XenServer 6.2 or 6.5. You need to install the VM driver as well as the XenServer drivers separately.
The GPU version can be run either standalone and with support for not just the GRID, but a number of different GPUs, or you can run it as a VM under XenServer in which case you'd want to install the XenServer portion so you can properly assign GPU passthrough to the VM. All the downloads are are at
http://www.nvidia.com/Download/index.aspx?lang=en-us. The combinations of what drivers are needed are a bit confusing.
As to preference, it all depends. Id you want a dedicated, guaranteed amount of GPU available to each desktop, the vGPU option might be a better bet. If you want anyone to be able to take advantage of a GPU with more power available when not so busy or less if so, then GPU passthrough may be a better fit. With multiple engiens to assign, you can even have a combination of the two. As to AutoCAD specifically, it works fine both ways. See also for sure this specific article on AutoDesk applications for GPU passthrough at http://blogs.citrix.com/2014/05/28/configuring-virtualised-autodesk-and-similar-applications-including-xenapp-gpu-sharing-and-tools-for-sanity-checking-directxopengl-usage/?_ga=1.146046580.1558783495.1433165527 (thanks for that article, Rachel!).
How many instances you can support will again depend in the case of VGPUs, how you carve up the allocations, or in the case of GPU passthrough, on what different users all might be doing at the same time.
Best regards,
嗨...
您有两种不同的选择。
您可以以直通模式（独立或作为XenApp安装）运行Windows服务器，在这种情况下，它需要驱动程序作为GPU（NVGRID系列），或者如果作为VM桌面运行，每个使用vGPU，它需要一个不同的驱动程序，
将与XenServer 6.2或6.5的下拉菜单部分（NVIDIA vGPU）中的XenServer捆绑在一起。
您需要单独安装VM驱动程序以及XenServer驱动程序。
GPU版本可以独立运行，不仅支持GRID，还支持许多不同的GPU，或者您可以在XenServer下将其作为VM运行，在这种情况下，您需要安装XenServer部分以便正确运行
将GPU直通分配给VM。
所有下载都在
http://www.nvidia.com/Download/index.aspx?lang=en-us。
需要哪些驱动程序的组合有点令人困惑。
至于偏好，一切都取决于。
如果你想要为每个桌面提供专用的，有保证的GPU数量，那么vGPU选项可能是更好的选择。
如果你希望任何人能够在不那么繁忙或更少的时候利用具有更多可用功率的GPU，那么GPU直通可能更适合。
有多个工程师可以分配，你甚至可以将两者结合起来。
至于AutoCAD，它的工作方式很好。
另请参阅http://blogs.citrix.com/2014/05/28/configuring-virtualised-autodesk-and-similar-applications-including-xenapp-gpu-sharing-中有关GPU直通AutoDesk应用程序的特定文章。
and-tools-for-sanity-checking-directxopengl-usage /？_ ga = 1.146046580.1558783495.1433165527（感谢那篇文章，Rachel！）。
您可以支持多少个实例将再次依赖于VGPU，如何分配分配，或者在GPU直通的情况下，依赖于不同用户可能同时执行的操作。
最好的祝福，
以上来自于谷歌翻译
以下为原文
Hi...
You have two different options. You can run the Windows server in passthrough mode (standalone or as a XenApp instalnce) in which case it will need a driver as a GPU (NVGRID Series) or if run as VM desktops using a vGPU for each, it needs a different driver that will be bundled with the XenServer in that dropdown menu section (NVIDIA vGPU) for XenServer 6.2 or 6.5. You need to install the VM driver as well as the XenServer drivers separately.
The GPU version can be run either standalone and with support for not just the GRID, but a number of different GPUs, or you can run it as a VM under XenServer in which case you'd want to install the XenServer portion so you can properly assign GPU passthrough to the VM. All the downloads are are at
http://www.nvidia.com/Download/index.aspx?lang=en-us. The combinations of what drivers are needed are a bit confusing.
As to preference, it all depends. Id you want a dedicated, guaranteed amount of GPU available to each desktop, the vGPU option might be a better bet. If you want anyone to be able to take advantage of a GPU with more power available when not so busy or less if so, then GPU passthrough may be a better fit. With multiple engiens to assign, you can even have a combination of the two. As to AutoCAD specifically, it works fine both ways. See also for sure this specific article on AutoDesk applications for GPU passthrough at http://blogs.citrix.com/2014/05/28/configuring-virtualised-autodesk-and-similar-applications-including-xenapp-gpu-sharing-and-tools-for-sanity-checking-directxopengl-usage/?_ga=1.146046580.1558783495.1433165527 (thanks for that article, Rachel!).
How many instances you can support will again depend in the case of VGPUs, how you carve up the allocations, or in the case of GPU passthrough, on what different users all might be doing at the same time.
Best regards,
举报
举报
