What are some use cases that require strong hardware?
Posted by bluz1n@reddit | learnprogramming | View on Reddit | 14 comments
Hey everyone,
I was talking with a friend about this, and it turns out that often we don't use all the hardware power we can get on our jobs. I'm a Golang developer and he is a Data Engineer, so our work mostly run on cloud (and mine specifically can be local and super lightweight).
So what are some use cases that requires a very strong and local processor/GPU/RAM, aside from 3D graphics and ML related stuff?
razorree@reddit
working on a big code, like monolith with JVM for example, running e2e tests were you spin a few test containers (like DB etc.), using IDE, DB browser, Postman and many tabs in browsers, Slack etc. and 32GB RAM is required. (better CPU, faster compile times and executions).
Comprehensive_Mud803@reddit
Generally, compute-intensive tasks. You already mentioned 3D and ML, but there’s more:
ha1zum@reddit
I forgot about navigation. GIS is still very huge and of course there are recent advancements in video mapping with drone footage and such.
UpperSun16@reddit
honestly? just trying to run Slack and three Chrome windows at the same time in 2026. that’s the real hardware stress test😂
Reasonable_Ad1226@reddit
Why bother asking? You clearly don’t know enough to make use of the answer…
d-k-Brazz@reddit
Large codebases
If you are working on a big company having millions lines of code their projects it will require significant CPU effort to compile it
Your IDE will require both cpu and memory to index the entire codebase so you could leverage IDE toolkit
Depending on what your software is doing, you may need resources to run it locally, especially if your code is operating with gigantic pieces of sample data
PhilNEvo@reddit
Physics simulation, I guess that's going to be related to 3D graphics, but even if you don't necessarily want to 3D render it, if you want to model stuff that either has high complexity or a lot of small parts.
There's also generally big data stuff-- Even if you just want to track and sample a bunch of data, let's say weather. Allegedly one organization produce almost 300 terabytes of data per day.
There's also academic research that tries to find solutions to exponential or factorial problems. For example looking for optimal/minimal static sorting networks using comparators.
JohnBrownsErection@reddit
Digital signal processing.
I used to write quite a bit of music using digital instruments and anything of any decent complexity would require the hardware to handle it, particularly with the beefier VSTs.
Aggressive_Ad_5454@reddit
I have lots of cores and RAM so I can run several VMs on my desktop machine. That's an important developer use case for many situations.
The GPU doesn't make much difference for me. But people doing video need it.
ha1zum@reddit
Local AI training, 3D graphics (game development, VFX, physics simulations, CAD engine development), high-res video processing, device simulations (mobile apps development and testing), OS virtualization (distributed backend development, network simulations)
dkarlovi@reddit
Video processing only for final render, you should always use proxies for the editing.
Zesher_@reddit
I worked on an iOS app at a major company, the app was massive, and it could take upwards of 5 minutes to compile and run. Then oops, something doesn't look right, gotta tweak something, well that's another 5 minutes. Tweak something else, another 5 minutes. Then a meeting, then something else, etc , and a small change takes half a day.
I convinced management to buy us better computers by showing how much money they would save with how much time we were wasting just sitting around vs how much they were paying us per hour.
So sometimes it's as simple is you can compile faster and that makes you more efficient.
plastikmissile@reddit
Basically anything that requires a ton of math being done quickly. Examples of that include:
AI and ML
3D graphics
Physics simulation
Hardware emulation
Breaking cryptography
Status_Branch863@reddit
Game dev here and yeah you definitely need beefy local machine when working with large game projects - compiling shaders and building assets can take forever on weak hardware