DEV Community

Discussion on: How much do hardware parts matter?

Collapse
iamschulz profile image
I am Schulz

Swapping only occurs if your RAM is full. Most install guides suggest deactivating swap altogether these days.
When working with heavy node tasks, a high cpu clock speed is important, because node doesn't really do multithreading. More cores come into play when working with docker.
Your GPU is mostly not being hit. Apart from training AI models, development isn't gpu-intensive.
The most important piece of hardware for me is a fast SSD. That actually speeds up the times needed for file operations - which is 90% of what I do day to day.

So yeah, your virtualbox hits hard on your CPU, but you're fine with everything else.

Collapse
stereoplegic profile image
Mike Bybee

I disagree. RAM makes a big difference in a virtualized OS (especially Win10 in Virtualbox - easily testable by increasing allocated guest RAM).

You can gain performance at the cost of some features being harder to configure (shared folders, clipboard, though both can still be accomplished) by using KVM (Kernel Virtual Machine, one of my favorite Linux features) instead of VBox, as it is closer to bare metal.

Collapse
patarapolw profile image
Pacharapol Withayasakpunt Author

I normally allocate 4GB to Windows on VirtualBox, on another MacAir (around 2018, I think).

And when running PowerBI, it still lags.

Collapse
patarapolw profile image
Pacharapol Withayasakpunt Author • Edited on

When working with heavy node tasks, a high cpu clock speed is important, because node doesn't really do multithreading. More cores come into play when working with docker.

Yes, I also use Node.js and Docker.

Thanks for explanation, though.

I don't yet know much about AI models, but I am currently learning about Decision Tree Analysis, Clustering, and Association Mining. But indeed, I heard about GPU requirements from Deep Learning, on YouTube.

I also tempted to try DaVinci Resolve (VDO editing).