GPU for non gamers

Off-Topic
1

i am still learning c and python as both are entry level and easy languages.

so, i am going to buy a new laptop soon.

i am confused choosing between AMD Ryzen™ 7 5800X, 32GB DDR4 Ram with 512GB NVME SSD
or
AMD Ryzen™ 7 5800X, 8GB DDR4 Ram with 256GB NVME SSD with RTX 3060

then i looked up for what is gpu and why it does.

then i found some use cases:

  1. Content Creation (photoshop , video shop etc)
  2. Web (like netflix or prime 4k streaming)
  3. Visualization or virtualization for AI or ML programmers i guess.
  4. Bitcoin Mining.
  5. Of course for gaming

but i dont do conetent creation nor web streaming ( Torrent long live) or any other above.

obviously , i should go for first configuration which is 32gb ram, with inbuilt or integrated graphics card.

what are your thoughts?
and
1) What are other use cases you know where GPU is useful for non gamers or for programmers?
2) Does Rocky LInux support NVidia RTX 30** graphics cards?

1 Like
2

Back in the 1990s computer manufacturers used to build both CPUs (you know what that is, I am sure) and math co-processors (a separate designated processor for higher-end mathematics).

The industry stopped making them, and then eventually got back to it, but in a different style. Now the math-coprocessor is essentially built into your GPU (Graphics Processing Unit), which also enables you to do math crunching by offloading those calculation responsibilities to the GPU.

The GPU is a video chipset, but so robust you can also take onthose calculations, previously described to the GPU. This is a Win-Win situation for corporations that need the ability to do high end math, and quickly. There is even application of GPUs to do the same thing but within virtualization environments; which is really darned cool!

If you are just starting off learning coding (ie. programming) then don’t go hog-wild over the GPU, unless you are also gaming. As for GPU support, the question is appropriate if you are looking to install the video drivers, as opposed to GPU drivers, which is a very thin line in terms of difference but remember GPUs can be used both as video cards as well as additional math-coprocessor capabilities.

If you are not doing photshop, nor crypto-currency mining, nor high end 3D object graphics design then you really do not need to go hog-wild with an expensive GPU.

GPUs are natively compatible for video, but if you are testing your video creations {3D object graphics}, then you want to get the more enhanced GPU.

You are on the right track with your questions.
Good luck.

3 Likes
3

Eh, not exactly.
First of all, x87 math coprocessors go all the way back to the 70’s, when Intel bought them from AMD and put their sticker on them. Yes, this odd relationship between Intel and AMD started all the way back then.

Math co-processors didn’t go away. They got integrated into the CPU. In fact, this was the main difference between a 486 “SX” and “DX” – the “DX” had an integrated math coprocessor! However, the “SX” still had an available add-on math coprocessor, the 487SX, which was, in fact, just a 486DX. Upon install of the “coprocessor”, the main processor would deactivate entirely. Yes, there shall be no SMP for you!

Anyway, similar thing is happening with GPUs in the last several years. They are no longer exclusively external parts, but are starting to be integrated into the same package as the CPU.

Now depending on what you are doing, the integrated version can actually be much better than an external card. Reason being shared resources. It takes a lot of time to pass data between the CPU and an external GPU, but nearly nothing if the GPU is integrated.

HOWEVER, external GPUs are much more flexible when it comes to thermals, so depending on the workload, either could be better. But in a laptop?

Well the way I figure it, a laptop really is a somewhat limited form factor, mainly because you have limits to the amount of power it can or should draw based on battery life, and also on how much heat it can release based on being physically smaller. Combined with the fact that the Ryzen 7 is a pretty solid part, I’d avoid additional external parts unless you really really need it, like if you are a game developer.

Your use cases 1-3 will be served very well by the Ryzen 7. Use cases 4-5 are pointless. Performance of the Ryzen 7 will probably be better under RL than the external anyway, since AMD puts everything into open source graphics, and nvidia is basically “for windows”.

For me, the choice is easy – more RAM and more SSD wins over something that won’t do anything useful.

EDIT: In case you’re wondering what kind of problems might come out of trying to nvidia… read this thread; Trouble in River City -- Rocky Linux 8.5 Rollover Blew Up - #16 by desercat

3 Likes
4

After going through all points, i suggest you stick with your option one with 32 GB RAM due to following points:

  1. Now days each software require big amount of memory and if you don’t have enough memory then you will struggle with slow response and you are also limited with VM.
  2. It’s not advisable to do crypto mining with your laptop because will not getting any cost based benefit.