I’ll wait for the Julian Assange review.
What the fuck is wrong with this timeline.
Do the amish accept atheists?
I can’t see the context because it’s freaking X, but I bet that’s in reference to local ML hosting.
There’s a big movement to get away from corporate AI, and I don’t need to explain the importance of that to the Lemmy crowd. But Nvidia is indeed artificially crippling consumer VRAM to stop them from being used for that too much, and protect their enterprise GPU market.
The most bizzare thing is that AMD is inexplicably complicit even though they have like zero market share in that space. 48GB 7900s (and so on) would have obliterated Nvidia and sold like hotcakes, much less actually using their modular memory controller architecture… But no? They restricted their OEMs from doing that because they… Don’t want money, I guess.
AMD is much of a scum as Nvidia is or Intel was, that’s why DeepSeek is something that came from China and you would need a new player completely outside of the current chain.
I don’t see how that’s relevant, Deepseek was trained on, and is served on, Nvidia hardware.
And while I don’t disagree about AMD gouging, if AMD was to act like “scum,” they would screw over Nvidia’s pricing strategy.
But… they don’t. And lose for it.
It’s some combination of ignorance, corporate stupidity and straight up collusion, but it’s also the opposite of greed.
That headline is so stupid that I refuse to read the article
But what does Ja Rule think?
An infamous former U.S. National Security Agency (NSA) contractor and whistleblower has unexpectedly shared his opinion on the state of the graphics card market.
Man who did big cool thing once also has opinions on unrelated thing, news at 11.
Edward Snowdon reads a spec sheet
So now we care what Edward Snowden says about vram? We need him to tell us that it should be 24 gigs?
What is even this ?
Kamoulox
Was this written by AI? The headline word salad contains all the buzzwords.
Edward Snowden PILEDRIVES the Nvidia RTX 50 series into a crowded bitcoin farm
“Trash fuckin cuck card kys”
Don’t that distract you from that fact that 2025 Edward Snowden threw Nvidia’s RTX 50 series off Hell In A Cell and plummeted 16 ft through an announcer’s table.
Whistleblows on poor performance is actually insane lol
You’re about 30 years too late, but I’m glad you’re finally here
Every time I see a headline that contains the word “slams,” I want to slam my head on the table
And the user who posted it. I really wish there was a simple way to block sensational posts from my feed.
Thanks for giving me the idea. BRB, making a keyword filter for the words “slams” and “slammed”.
But then you lose out on a bunch of wrestling posts. And naked wrestling posts.
Careful now, you might end up with few news posts
Don’t forget blasted and clapped back.
Yeah that does seem a bit weird, my four year old RX 6800 has 16 GB VRAM already
Yes but that’s from a company that builds decent cards at an acceptable price, and not from the HypeLords of the Lesser Return on Investment Unless You’re a Farm.
(Getting ready for the downvotes!..)
Edward Snowden doing GPU reviews? This timeline is becoming weirder every day.
“Whistleblows” as if he’s some kind of NVIDIA insider.
Intel Insider now that would’ve made for great whistleblowing headlines.
Swear next he’s gonna review hentai games
“Some hentai games are good” -Edward Snowden
Note that this is from 2003
I’ll keep believing this is a theonion post
Legitimately thought this was a hard-drive.net post
I bet he just wants a card to self host models and not give companies his data, but the amount of vram is indeed ridiculous.
Exactly, I’m in the same situation now and the 8GB in those cheaper cards don’t even let you run a 13B model. I’m trying to research if I can run a 13B one on a 3060 with 12 GB.
You can. I’m running a 14B deepseek model on mine. It achieves 28 t/s.
You need a pretty large context window to fit all the reasoning, ollama forces 2048 by default and more uses more memory
Oh nice, that’s faster than I imagined.
I also have a 3060, can you detail which framework (sglang, ollama, etc) you are using and how you got that speed? i’m having trouble reaching that level of performance. Thx
Ollama, latest version. I have it setup with Open-WebUI (though that shouldn’t matter). The 14B is around 9GB, which easily fits in the 12GB.
I’m repeating the 28 t/s from memory, but even if I’m wrong it’s easily above 20.
Specifically, I’m running this model: https://ollama.com/library/deepseek-r1:14b-qwen-distill-q4_K_M
Edit: I confirmed I do get 27.9 t/s, using default ollama settings.
Ty. I’ll try ollama with the Q-4-M quantization. I wouldn’t expect to see a difference between ollama and SGlang.
Thanks for the additional information, that helped me to decide to get the 3060 12G instead of the 4060 8G. They have almost the same price but from what I gather when it comes to my use cases the 3060 12G seems to fit better even though it is a generation older. The memory bus is wider and it has more VRAM. Both video editing and the smaller LLMs should be working well enough.
I’m running deepseek-r1:14b on a 12GB rx6700. It just about fits in memory and is pretty fast.
Does he work for Nvidia? Seems out of character for him.
