Nvidia's Chinese website continues to list workstation-grade RTX 6000 Ada solutions. The RTX 6000 Ada is the company's most powerful graphic card and is subject to export restrictions to China (and some other countries), according to the latest U.S. trade rules. In fact, with a compact blower cooler and 48 GB of memory, the RTX 6000 Ada is a very good fit for artificial intelligence training. In contrast, Nvidia and its partners no longer sell the GeForce RTX 4090 graphics card in China, VideoCardz noticed, and the company has removed any mention of the consumer product from its Chinese website.
Nvidia's GeForce RTX 4090 — one of the best graphics cards around — is based on the AD102 graphics processing unit and has a total processing performance score of 5,280 (based on its FP8 Tensor FLOPS performance of 660 TFLOPS), which makes it an export-licensable item (as its TPP is higher than 4,800). To ship GeForce RTX 4090 products to China, Nvidia and its partners now have to obtain an export license from the U.S. Department of Commerce. Such license applications are reviewed with a presumption of denial, so it looks like Nvidia would rather not sell its GeForce RTX 4090 in China.
With the above in mind, it is strange that the company continues to offer its RTX 6000 Ada Generation graphics board for professionals in China. This solution features the AD102 graphics processing unit with 18,176 CUDA cores enabled; its total processing performance score is 5,828 (based on its 728.5 FP8 TFLOPS performance without sparsity). In fact, this graphics card carries 48 GB of memory, so it is actually a better fit for AI training and inference than the consumer-oriented GeForce RTX 4090.
|GPU||RTX 6000 Ada Generation||GeForce RTX 4090|
|Architecture | GPU||Ada Lovelace | AD102||Ada Lovelace | AD102|
|Memory||48 GB GDDR6 w/ ECC||24 GB GDDR6X|
|Total Processing Power (FP16/BF16)||5,828||5,280|
|Memory Bandwidth||960 GB/s||1008 GB/s|
|INT8 I FP8 Tensor||728.5 I 1457 TFLOPS||660 I 1320 TFLOPS|
|BF16 I FP16 Tensor||91.06 TFLOPS||82.58 TFLOPS|
|FP32||91.06 TFLOPS||82.58 TFLOPS|
|FP64||1423 GFLOPS||1290 GFLOPS|
|L2 Cache||96 MB||72 MB|
|Form Factor||2-slot FHFL||3.5-slot|
|Interface||PCle Gen4 x16: 64 GB/s||PCle Gen4 x16: 64 GB/s|
What is perhaps more important is that Nvidia's RTX 6000 Ada Generation comes with a compact blower cooling system, which makes it easy to use in data center environments and which makes it useful both for AI training / inference jobs as well as high-performance computing, as it supports FP64 without constraints.
It is unclear why Nvidia still lists its RTX 6000 Ada Generation graphics card on its Chinese website. Perhaps the company intends to apply for an export license to keep selling this $6,800 product in the the People's Republic, or maybe the company's partners have plenty of those cards in China already and therefore will be able to continue selling stock for some time.
Stay on the Cutting Edge
Join the experts who read Tom's Hardware for the inside track on enthusiast PC tech news — and have for over 25 years. We'll send breaking news and in-depth reviews of CPUs, GPUs, AI, maker hardware and more straight to your inbox.
Anton Shilov is a Freelance News Writer at Tom’s Hardware US. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.
It seems like this AI world wide power struggle is a frighting reality.Reply
One can see the good to man kind but in reality so was modern medicine where some meds are a $1000 a pill. Why $$$$
Another layer to add to man kind. $$$$
Funny all I can think about is please don't put a AI controlled CPU on my toaster.
But it is scarry that most of everything we do is believe who we a dealing with is a real person on the other side of the phone. But in reality that's just our own inner circle of family/ coworkers/ freeway traffic and grocery shopping the whole rest of the world for all we will know can be ran by Ai.
Skynet is learning but who's in charge of Skynet is in charge.
Yeah, this really gonna work. Just like Steam Deck is available everywhere in the world despite haven't "officially" been released outside of only few countries. Those who want/need this stuff can certainly get it despite being just common folks and here we are talking about multi-billion dollar corporations or perhaps even governments. 😉Reply
So technically this should reduce the price of 4090 worldwide, but Huang will some how make sure it wont happen.Reply
Except that it won't. The Chinese will use a third party country, that nVidia is still allowed to ship to, to get them, in the same way that Russia is circumventing the sanctions.vinay2070 said:So technically this should reduce the price of 4090 worldwide, but Huang will some how make sure it wont happen.
GPURTX 6000 Ada GenerationGeForce RTX 4090Memory48 GB GDDR6 w/ ECC24 GB GDDR6XTotal Processing Power (FP16/BF16)5,8285,280Performance Density9.578.66CUDA Cores18,17616,384FP3291.06 TFLOPS82.58 TFLOPSL2 Cache96 MB72 MBPower300W450WForm Factor2-slot FHFL3.5-slotInterfacePCle Gen4 x16: 64 GB/sPCle Gen4 x16: 64 GB/sReply
I find it interesting the "Total Processing Power" is virtually identical, yet it has 10.9% more fp32 TFLOPS. The additional performance is very interesting, in light of using just 2/3rds of the power.
Also, that extra 24 MB of L2 cache looks pretty nice.
Finally, the conventional way to specify PCIe bandwidth is uni-directional, in which case it should be just 32 GB/s. I'm assuming these numbers were taken from Nvidia, who seems to like inflating their specs at every opportunity. However, the reason why uni-dir is better is that you almost never have perfectly symmetrical data movement in both directions. Anyway, the more useful number is the uni-dir number.
This will increase prices for Chinese buyers, due to having to pay middle men + more shipping & duties. Higher prices usually translates into lower demand. As a result, we should expect to see lower sales volumes going to China, which theoretically could impact pricing & availability elsewhere in the world.Joseph_138 said:Except that it won't. The Chinese will use a third party country, that nVidia is still allowed to ship to, to get them,
That also explains why I think even leaky sanctions aren't necessarily pointless, depending on the goal.
Your family and coworkers can be using AI agents. This could range from something like a Siri assistant, who's scheduling a meeting or dinner reservation on behalf of them, to a Microsoft Office Copilot-written email they send you.stonecarver said:But in reality that's just our own inner circle of family/ coworkers/ freeway traffic and grocery shopping the whole rest of the world for all we will know can be ran by Ai.
Freeway traffic? Are you serious? Have you not seen the news about all the collisions & even deaths caused by Tesla's autopilot? Just because you see a person sitting behind the steering wheel doesn't mean they're the one driving!
I was thinking about the guy three lanes over flipping me off. But your right darn Tesla's.bit_user said:Freeway traffic? Are you serious?
You know my wife his been being nicer to me on her texts. :rofl: Maybe Ai has it's place.bit_user said:Your family and coworkers can be using AI agents. This could range from something like a Siri assistant, who's scheduling a meeting or dinner reservation on behalf of them
I was wondering why we stopped text like this " from Her" Drop dead " me back" you first..........😈
What's stopping anyone just going to buy them off from taiwan?Reply