Unable to configure my laptop with gpu quadra m1000m having windows7

Part 1 (2017)
1

I have an hp zbook 15 g3 with nvidia quadra m1000m gpu.
I tried with latest version of the nvidia driver downloaded from below link but installer fails to install Nvidia WMI 2.30
Latest Nvidia Driver: Latest Nvidia Driver
Latest driver version : 23.21.13.8508

I contacted nvidia customer care and they provided me some info to remove some thrid party softwares “LANDESK” and then install the nvidia driver. But that didnt wok either as WMI install still fails.

FYI …
Using with older drivers create some other issues.
Older nvidia driver version : 21.21.13.7651
CUDA Version: 8.0
Runing a simple two line commands causes blue screen

import pygpu
pygpu.test()

I am stuck in using my laptop gpu. Help would be greatly appreciated.

2

To share what I have done on my desktop pc with gpu GeForce GTX 1060 having windows 7.

I bought a Nvidia 1060 6GB as an addon for my desktop pc.
OS: Windows 7 Enterprise 64 bit
Hardware: HP EliteDesk 800 G2 SFF

Following the setup instrunctions in https://github.com/philferriere/dlwin, I was able to go on with the course materials.

d:\toolkits.win\anaconda3-4.4.0
d:\toolkits.win\cuda-8.0.61

Running the same test commands gave me the following result.

pygpu is installed in D:\toolkits.win\anaconda3-4.4.0\envs\dlwin36\lib\site-packages\pygpu
NumPy version 1.13.3
NumPy relaxed strides checking option: True
NumPy is installed in D:\toolkits.win\anaconda3-4.4.0\envs\dlwin36\lib\site-packages\numpy
Python version 3.6.2 |Continuum Analytics, Inc.| (default, Jul 20 2017, 12:30:02) [MSC v.1900 64 bit (AMD64)]
nose version 1.3.7
*** Testing for GeForce GTX 1060 6GB
mpi4py found: False
......................................................................................................
================================================
pygpu.gpuarray.GpuArrayException: b'NCCL is not available on plaforms other than linux.'

To install theano 1.0.0rc1, I tried:

conda install -c mila-udem/label/pre theano

To verify,

>set THEANO_FLAGS=%THEANO_FLAGS_GPU%
>python cpu_gpu_test.py

Looping 1000 times took 0.473095 seconds
Result is [ 1.23178029  1.61879349  1.52278066 ...,  2.20771813  2.29967761  1.62323296]
Used the gpu

Which looked good to me.