Z by HP Data Science Software Stack — the answer to your set-up hassle

4 min readApr 22, 2021

Introduction

I wanted to do this blog post for all Data Scientists out there that love the science, but not the set-up part of it.

Indeed, many people like to do everything themselves, and this also includes setting up their Data Science Environment. They like installing each package one by one, check out new versions, look out for new improved options.

However, I am not one of these people. Are you?

If not, in this article I am going to present you with the Holy Grail of the Data Science Environment set-up. Why can I tell you about this? Because months on end I went through issues on setting up my perfect Data Science Environment, up to a point where I wasn’t even doing much Data Science at all, to be honest. However, I was deep into articles on CLI, Windows set up, WSL2, command errors, dependencies errors, driver incompatibilities … all that cute package and all.

No more. The hassle is over.

How it was

I would like to layout a bit of the context I was in. If you aren’t interested in this part, just jump to the next chapter, where I tell you how I got rid of all my problems.

I started with a Z8 G4 Workstation from HP, with some pretty powerful features I would say: a 96 GB of memory, 2 CPUs, and a sweet NVIDIA Quadro RTX 8000. This dragon was ready to train some models and process some data for sure, and I was super psyched about it.

It took me around 2–3 days to properly install my usual software. Some of them were:

Anaconda
PyCharm
Git
PyTorch (and many more other libraries that don’t come pre-installed with Anaconda)
XGBoost
R Studio
Nvidia CUDA

Then I proceeded to install RAPIDS — how to have an NVIDIA graphics unit and NOT install RAPIDS?

And here I hit the wall, and hard I would say. RAPIDS works only on Linux and I was working on Windows 10. Hence, I proceeded to update my windows (which took a few hours), installing WSL2 and then setting up RAPIDS (for this process I did a detailed article with code here). All this took a bonus of another few days. Is anybody counting?😪

However, there was an issue: mounting the container. Long story short, when closing the RAPIDS environment after your work is done everything gets purged: data, notebooks, even installed libraries. This isn’t useful when you’re trying to work on a long term project.

Then I proceeded to move to Linux; from my research, this is the OS that Data Scientists are starting to use anyways. However, setting up the environment STILL proved to be a hassle, starting aaaaaaaaaall over again, doing tests to see that everything works how it should, other super specific errors…

I won’t lie: I was a bit worn out.

Then, I found out about the Z by HP Data Science Software Stack.

The Z by HP Data Science Software Preload

It is unreal. It is perfect. It is quick. It doesn’t require your attention. And after it’s done everything works perfectly and you can start Data Sciencin’ instantly.

As of now, the Stack is available for HP Z4, HP Z6, HP Z8 desktop workstations and is planned for the end of December on the HP ZBook Studio G7 and HP ZBook Create G7 (but it will be extended in 2021). And … it comes with everything.

The Data Science Software Stack comes with:

The latest Python & R version
Preloaded environments like Anaconda, PyCharm, Visual Studio
GIT
Data Science packages like PyTorch, TensorFlow, XGBoost
RAPIDS: with an Anaconda environment where you can easily switch between base and rapids
Docker Container and Nvidia Drivers
etc.