Hands on with NVIDIA P6 and UCS B200 M5

Just got access to a new UCS B200 M5 blade!  My goal is to create a tensorflow lab on it.  Let’s get cracking!

It was installed with CentOS 7.5

cat /etc/redhat-release

Let’s make sure there is indeed a GPU:

Ok, now we need to get some drivers. We go to NVIDIA’s page, fill out the form and get some drivers for RHEL7.

While waiting for downloads, we make it so we can sudo without a password. Run sudo visudo and edit these lines:

Now let’s install the VNC server and some other packages we’ll need.  This will give us development tools and a remote desktop

Now let’s attach to it… hmm.  we can’t.  Is selinux running?

yep.  Let’s turn that off for now.  We don’t need this.  Use sudo to modify /etc/sysconfig/selinux

Have to reboot.  But first let’s install the driver:

edit /etc/default/grub to disable the nouvea driver.

load the new grub file

Add the nouvea driver to the blacklist by appending to (or creating in my case) /etc/modprobe.d/blacklist.conf

Back up the old stuff and make the new initrd

Now we reboot.

I installed the the 32 bit compatible libraries because diskspace is cheap and time is short.

CUDA Libraries

We want tensorflow with the CUDA libraries.  It makes tensorflow fast!  We get it by navigating to their page.  I downloaded the runtime one.

I answer the questions as follows:

Since all went well you will see the following output

(or at least something similar)

Now lets get the environment setup.  Append to ~/.bash_profile

(I’m using 9.2 as this is the version of cuda I’ve installed, it may be different when you install so change as updates come available.)

cuDNN Library

To download the cuDNN libraries you need to have an NVIDIA developer account.  You’ll have to login to the download site and download the Linux version

Here we are cuDNN v 7.2.1 with CUDA 9.2 as that is the library we used.


Installing Tensorflow

Install pip

Now we can install tensorflow:








Installing NVIDIA Drivers on RHEL or CentOS 7