A GPU from 2021 is better than anything I have and unlike nvidia they aren’t complicit in genocide.
Chinese capitalists are still capitalists.
You do realize they’re only going to get better.
better than i expected, actually. software issues and optimization aside, they might be a couple of gens from pretty good. i wonder how they will be able to scale up production now.
as i remember, they are using an old process in a new way to get euv performance, but that’s cutting on yields, right?
I’m guessing it’ll take a year or two for that to get ironed out. The nice thing about the tau folding process Huawei came up with is that it’s complimentary with euv. So, once Chinese companies do master it, their chips will be that much better.
The rtx 3060 is a fine gpu in 2026. Can it play modern aaa games in 4k ultra setting? No. But people just want cheap gpu’s back no one wants to pay half a grand for an entry level card.
Im still gaming on a 3060 and can even the most modern games at 1080p low/med settings.
Man I’d love to be able to mesh around with the gaming scene in China.
What’s stopping you?
Knowledge mostly the rest is probably time but mostly knowledge. I know just enough tech wise but not nearly enough to be helpful.
Competition as last
Not really. You’d be better off buying an Arc card.
Is 30,000 a lot for China? I feel like it might not be.
Also this is kind of the bottom line: “The LX 7G100 remains a poor buy for anyone chasing performance per dollar…”
I expect the costs will come down once they scale up production.
My feelings:
-
Neat.
-
Future is interesting.
-
But proof is not in the puddin’ yet.
It’s easy to say “they’ll scale it up. They’ll get to optimizing the software,” but I’ve seen waaay too many hardware makers mess up that step, and fade into the vast graveyard of their peers.
Like, does anyone remember Centaur’s exciting, 8 core x86 CPUs, with a huge block for ML inference, from about a decade ago?
https://chipsandcheese.com/p/via-part-4-a-deep-dive-into-centaurs-last-cpu-core-cns
It was competitive with Zen 1, not even counting the accelerator.
The answer is “probably not.” Even the Wikichip Fuse writeups on the architecture are gone.
I mean China has a proven track record of scaling stuff up once they decide to do something. Look at solar and EVs as two examples. Having sovereign supply chains for computing is obviously recognized as a top priority in China now, so there is every reason to expect that a lot of resources will be poured into making that happen at state level.
That’s a lot of vagueness though.
I’m interested in something physical. Something I could buy, that could realistically upgrade an Arc B580 in (say) indie games, or KCDII, or llama.cpp.
This is nowhere close yet.
And I dont doubt they’re getting investment, but I’m just skeptical because I’ve seen this story a hundred times before, even in China. And even then, what’s promised to be general hardware often evolves into something for a very nich need. As a recent example, see Tenstorrent.
I mean all we can do is wait and see, but there’s nothing magical about this tech, and the geopolitical situation today is very different from what it was in the past. It’s not like Chinese people are too dumb to build a GPU, it’s that there was no real reason to push this tech hard before when you could just buy these chips from NVIDIA. Now that the US is actively trying to cut China off, there’s a reason to make this stuff work. Just look at how rapidly Huawei phones improved and how much progress Huawei is making with their chips in just past few years. Why shouldn’t this happen with other types of chips going forward?
-
deleted by creator
