问题描述
我在 Intel Core i7-8700 CPU 上运行 Windows 10，配备 16 GB RAM、1 TB HDD 和专用 NVIDIA GeForce GTX 1070 显卡.
I am running Windows 10 on Intel Core i7-8700 CPU with 16 GB RAM, 1 TB HDD and dedicated NVIDIA GeForce GTX 1070 graphics card.
我计划启动由我的 Windows 10 PC 托管的 3 个 Ubuntu 实例.Ubuntu 将运行分布式 Tensorflow (tensorflow-gpu) 代码，该代码将使用 GPU 来训练神经网络.(值得一提的是，已经我已经在 Windows 上尝试过设置但失败了)
I plan to launch 3 Ubuntu instances hosted by my Windows 10 PC.The Ubuntus will be running Distributed Tensorflow (tensorflow-gpu) code, that will using GPU for training a Neural Network. (to mention, already I've tried the setup on Windows but failed)
问.我的 NVIDIA GPU 是否应该在这些虚拟机之间进行虚拟化?
如果是，那么是否需要进一步的配置来实现这一点?
If YES, then is there any further configurations required to make this happen?
如果没有，那么有没有建议为分布式 Tensorflow 构建这样的实验环境?
If NOT, then is there any suggestions to build such experimental environment for Distributed Tensorflow?
注意
我已阅读这篇文章说虚拟机无法通过主机GPU，特别是在 Windows 上用于 CUDA.但是是否有任何可用的最新信息，最好是来自 NVIDIA 方面的信息?
I have read this post saying VMs can not pass through host GPU, specifically on Windows for CUDA. But is there any recent information available, ideally from NVIDIA side?
任何人都可以分享一些操作方法，以便(可能)在 Windows 上的 ESXI 设置中虚拟化 GPU?就像有几个人在谈论这里有可能并且已经完成，虽然 未得到 NVIDIA 官方支持.
Can anyone share some how-to follow in order to (possibly) virtualize GPU inside ESXI setup on Windows? Like there are several people talking here that its possible and done, although not officially supported by NVIDIA.
或者，是否有人成功实施了在基于 Debian 的系统上推荐的 GPU 直通解决方案?
Alternatively, has anybody successfully implemented this suggested solution for GPU pass through on Debian-based system?
谢谢.
推荐答案
我会考虑 @jdehesa 的答案，因为现在似乎没有办法在 Windows 上为 Tensorflow 虚拟化 GPU.感谢@jdehesa
I would consider @jdehesa's answer as for now there seems no way to virtulize GPU on Windows for Tensorflow. Thanks to @jdehesa
这篇关于可以为分布式 Tensorflow 虚拟化 NVIDIA GeForce GTX 1070 显卡吗?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！