this post was submitted on 23 Nov 2023
1 points (100.0% liked)
AMD
25 readers
1 users here now
For all things AMD; come talk about Ryzen, Radeon, Threadripper, EPYC, rumors, reviews, news and more.
founded 11 months ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
is throwing $1k at gpu worth it?
Sure is if the alternative is spending over $2000 for a GPU with slightly more performance.
what is slightly more performance ?
0-30% faster for well over 100% more cost.
I mean if you're gonna include the games that run better on AMD in your range, might as well include the games that run better on Nvidia no? It's more like 0-200% faster.
All that aside the 4090 is in it's own league, no competitors what so ever. It's bad value yes but water is also wet, and it's price is currently inflated because of how it's insanely good at AI. 7900XTX is also poor value compared to literally any other RDNA3 GPU and unlike the 4090, it isn't a halo product, but it's pretty good if you have max 1k to spend on GPU.
Definitely. You speak the truth. Also, AMD cards suck in anything that isn't gaming - well, maybe an argument can be made for video editing - but, that's about it. Two areas of gpu use - it's good at and nothing else. For over $1k - well, I guess Americans can buy them for below that but anywhere else? No.
We've gone over this, that just isn't the case.
The 7900XTX is 4090 levels of fast in ML like Stable Diffusion, in video editing with Resolve/Premiere, in professional CAD like Soildworks.
It's just as good as a 4080/4090 in Photoshop, After Effects, Affinity Photo 2 and more.
I have no idea where you get this idea that it can't do anything except run games. It's just false.
Have you? I guess I wasn't here for that. I am wondering why your 'take' is the opposite of ppl who participate in the Machine Learning sub, Blender sub, Stable Diffusion etc. etc. - I guess they're lying or you know more than all of them?
I'll give you video editing even though the ppl at Puget Systems who spend their time doing benchmarks and comparing gpus for all that software recommend Nvidia cards. But, I digress.
For ML, it's news to me that AMD cards are just as good or you refer specifically to a 7900 XTX?
https://www.reddit.com/r/MachineLearning/comments/17kqexv/started_machine_learning_what_gpu_discussion/
https://www.reddit.com/r/MachineLearning/comments/qq21s6/d_why_does_amd_do_so_much_less_work_in_ai_than/
https://www.reddit.com/r/MachineLearning/comments/xjnozq/d_whats_the_word_on_amd_gpus_these_days/
Post your reply on the ML sub and if the majority agree with you, then I'll retract my assertions/assumptions.
Thanks.
I included all games. That's an average. There's no game that runs 200% faster on a 4090 over a 7900XTX. Is there any software at all that runs 200% faster on a 4090 over a 7900XTX?
But that isn't the alternative.
That's like saying buying a Ferrari is the only alternative to buying a Lamborghini.
The alternatives are any GPU less expensive than $1000. And that's the majority of GPUs.
I would not agree with that analogy but if you want to use cars, I would say it's like a Nissan GT-R is an alternative to a Ferrari. Most of the performance, half the price. And half the price means it's available to a much wider market.
Surely you can see $999 is a different segment to $2200.