Image: NVIDIA

Following this morning’s GeForce RTX 3080 Ti report, kopite7kimi has returned with fresh rumors regarding NVIDIA, which is purportedly prepping two additional Ampere graphics cards: the GeForce RTX 3060 and GeForce RTX 3050 Ti.

According to kopite7kimi, the GeForce RTX 3060 and GeForce RTX 3050 Ti leverage a new Ampere GPU codenamed GA106. These graphics cards will supposedly feature 3,840 CUDA Cores and 3,584 CUDA Cores, respectively.

In comparison, NVIDIA’s GeForce RTX 3060 Ti (which is already being sold, apparently), features 4,864 CUDA Cores.

The memory configuration of these graphics cards are unknown, but if its predecessors are anything to go by, the GeForce RTX 3060 should come with 6 GB of GDDR6.

kopite7kimi also shared a little bit about the process technology that NVIDIA will use for its Hopper architecture, which will presumably power the next-generation GeForce RTX 40 Series.

Echoing a claim made back in March, kopite7kimi has confirmed that two of NVIDIA’s Hopper-based GPUs, the GH100 (data center accelerator) and GH20X (GeForce) will be manufactured on TSMC’s 5 nm process.

Previous rumors suggested that NVIDIA could be planning an Ampere refresh on 7 nm.

Join the Conversation

24 Comments

  1. The rumour mill is working overtime it seems.

    On a more serious note, that amount of cuda cores between the 3050ti and 3060 is realy close to one another given that the 3060ti has like 1k more then the 3060.

  2. So going away from TSMC hasn’t payed off for nvidia.

    Not IBM, not Samsung, usmc was the other?

    BTW Samsung just announced its 5nmLP process, any chance nvidia could wait for the 5nmHP process?

    1. I doubt it because Samsung’s fab is lower quality than TMSC which is why nVidia got a huge discount.

      Samsung 5nm is questionable, especially compared to tsmc.

      The only reason they would be attempting 5nm (if this is true) would be because of power requirements. They need a node shrink to bring that down because ampere is absurdly power hungry.

  3. Doesn’t the title say 40 series. Are we already talking about that. Has Nvidia abandoned all hope of the 30xx series?
  4. Doesn’t the title say 40 series. Are we already talking about that. Has Nvidia abandoned all hope of the 30xx series?

    Since the 3000-series has hit retail, the 4000-series has almost certainly been in development for several years, and the process nodes mentioned are also reaching the point of being available to tape products out for test runs.

    It does seem a bit leakier than usual though, whatever implication that has :)

    So going away from TSMC hasn’t payed off for nvidia.

    Not IBM, not Samsung, usmc was the other?

    BTW Samsung just announced its 5nmLP process, any chance nvidia could wait for the 5nmHP process?

    I’m not quite ready to stick a fork in Samsung for GPUs in my mind, if only because I feel the need to leave some room for Nvidia’s potential shenanigans, i.e., holding back stock for some conspiracy-theory-grade reason. It’s not like Nvidia’s parts aren’t competitive, and in terms of their overall solution they still have AMD beat by a pretty decent margin, if only they could get some stock available.

    I’m also wondering if Nvidia isn’t going to use Samsung even more. Since Intel is pushing into the lower-end GPU game, it might make sense for Nvidia to use their available TSMC capacity for smaller, efficiency-focused parts where they can get the most performance possible at key TDPs, thus pushing their larger GPUs toward Samsung where a few extra watts (and I mean a few, not 50% more) aren’t quite as big of a deal.

    It’s also entirely possible that with Samsung producing Ampere as it is today successfully, that future runs might be better optimized from the application of lessons learned with Ampere. I’m sure that Samsung is absolutely ecstatic to get a piece of Nvidia’s pie and that any perceived shortcomings that the production of Ampere exposed are high on their list of issues to address sooner rather than later.

    I’ll also muse that Nvidia’s work with Samsung opens the door for other companies to follow suit, i.e. AMD and Intel. For AMD, Samsung would open up a tremendous amount of capacity add CPU production volume, which they’ll need if they’re going to truly square off with Intel’s marketshare. AMD can win every single benchmark but if they can only produce 1/20th of the product, then they can only sell 1/20th of the product (my made up numbers for illustration only). For Intel, while I doubt they’d use Samsung for CPUs, I think Samsung would be perfect for bigger GPU architectures. Again, once you jump into the TDP bracket of needing an external power connector (around 75w to start), extra wattage is only one consideration. In Intel’s case just having a working product on the market would be worth minor compromises.

    [last, apologies to @Stoly, I cued off your post, but this is just general conjecture, not a rebuttal or correction :) ]

  5. Doesn’t the title say 40 series. Are we already talking about that. Has Nvidia abandoned all hope of the 30xx series?

    No. You don’t think that companies develop new technology over the course of just 1.5-2 years, do you? We’ll probably be talking about GeForce Hopper in full force around this time next year. I think the window for a gaming Hopper release will open around summer of 2022. Since Samsung’s 8nm is more of a transitional process node for NVIDIA I don’t think it’s out of the realm of possibility that this will be a shorter video card generation this time around (14-20 months instead of 25 months).

  6. One of these days, I will buy a brand new GPU. Yes I will. Anytime now. Can’t be much longer now. Yep.
  7. I’m not quite ready to stick a fork in Samsung for GPUs in my mind, if only because I feel the need to leave some room for Nvidia’s potential shenanigans, i.e., holding back stock for some conspiracy-theory-grade reason. It’s not like Nvidia’s parts aren’t competitive, and in terms of their overall solution they still have AMD beat by a pretty decent margin, if only they could get some stock available.

    I’m also wondering if Nvidia isn’t going to use Samsung even more. Since Intel is pushing into the lower-end GPU game, it might make sense for Nvidia to use their available TSMC capacity for smaller, efficiency-focused parts where they can get the most performance possible at key TDPs, thus pushing their larger GPUs toward Samsung where a few extra watts (and I mean a few, not 50% more) aren’t quite as big of a deal.

    I think the opposite makes more sense. Using TSMC for high end parts and Samsung for efficiency focused parts, like say the tegra SoC.
    As of today, Samsung has no other process than the tailored for nvidia 8nm (which AFAIK is closer to 10nm than 7nm). Samsung has no experience building chips that big, hence the supply issues. So IMO it makes more sense to build smaller chips on Samsung and the bigger ones on TSMC.

    On a sidenote, remember the rumors that nvidia was going Intel for its 14nm and 10nm chips? That was supposed to give them the edge on manufacturing process, funny how things turned out…

  8. Doesn’t the title say 40 series. Are we already talking about that. Has Nvidia abandoned all hope of the 30xx series?

    Yeah… seems they are running a bit scared. Which is odd. Maybe they see how AMD has gained on Intel and are afraid of the same.

  9. Yeah… seems they are running a bit scared. Which is odd. Maybe they see how AMD has gained on Intel and are afraid of the same.

    It just seems odd to start leaking 4 thousand series GPU’s before they even have stock on shelves for the 3 thousand series GPU’s. That’s why I think Nvidia is flipping out and is going to focus on the next generation of GPU’s.

  10. It just seems odd to start leaking 4 thousand series GPU’s before they even have stock on shelves for the 3 thousand series GPU’s. That’s why I think Nvidia is flipping out and is going to focus on the next generation of GPU’s.

    Hmm.. Pull a Fermi 500 series? I guess it’s possible, but it certainly seems premature. AMD’s reviews haven’t even hit yet. I believe AMD will be competitive, but I don’t think it will be a bloodbath like 400 series Fermi vs Cypress was.

    I tend to think this is just more FUD being thrown out there to try to prevent people from jumping on AMD 6Ks, while nV tries to get their supply issues sorted out. Anything to stoke the Team Green fan club, and keep them from jumping ship while they wait for product to get delivered.

    Going a bit further, I think nV just released too early. Now they are trying to play catch up, and the rumor mill is one way to keep people engaged and get free news coverage while they are backpedaling trying to get logistics up to snuff.

  11. It just seems odd to start leaking 4 thousand series GPU’s before they even have stock on shelves for the 3 thousand series GPU’s. That’s why I think Nvidia is flipping out and is going to focus on the next generation of GPU’s.

    It actually isn’t all that uncommon for the rumor mill to start about the next architecture shortly after a new one is released.

  12. It actually isn’t all that uncommon for the rumor mill to start about the next architecture shortly after a new one is released.

    I agree… shortly after the new one is released. I would say that they haven’t COMPLETED releasing the current one.

  13. Using TSMC for high end parts and Samsung for efficiency focused parts, like say the tegra SoC.

    That’s the thing; Samsung 8nm doesn’t seem to be as efficient as TSMC 7nm.

    Not that you’d ever really be able to know the answer, but a smaller, more efficient node seems to make sense when the smaller parts are where the volume is. Not just thinking desktops here either, but also laptops. Everything from the MX line to the top entry-level part (currently 1650Ti) get sold as performance boosters, and that marketing bullet is becoming more and more reality.

    On the other hand, the larger parts are where the margins are, and they can absorb a part that runs a bit hotter. Not a lot hotter, but a bit. They’ll get derided for losing their usual perf/watt advantage over AMD but at least they’ll be competitive in terms of pure performance and will almost certainly retain their lead in consumer experience.

    Anything to stoke the Team Green fan club, and keep them from jumping ship while they wait for product to get delivered.

    Quite honestly, AMD is the reason that I’m not interested in buying AMD. While they’ve been focusing on catching up their hardware, they’ve managed to fall behind in every other way, or in some cases (drivers / software compatibility / release-day issues) just never make much headway.

    I don’t want them to stop trying and I do regularly evaluate them for my uses, and recommend them to others when they’re a good fit, but that’s also about as common as their marketshare would indicate. And I’ve recently been bit by them again.

    Going a bit further, I think nV just released too early. Now they are trying to play catch up, and the rumor mill is one way to keep people engaged and get free news coverage while they are backpedaling trying to get logistics up to snuff.

    Whatever the real reason, it’s become clear that they either messed up or they made a calculated decision to release with far too little stock. If the former, that’s a bit out of character as they’re usually pretty smart about that stuff; if the later, well, obviously it’s costing them, but people still want the GPUs, so perhaps it’s not costing them more than they expected.

  14. Not like most of us even had a chance of 30 series cards. For a 40s series to be mentioned already is interesting . I debsting to give AMD 6900 XT a shot even thougj i always been a nvidia user.
  15. Not like most of us even had a chance of 30 series cards. For a 40s series to be mentioned already is interesting . I debsting to give AMD 6900 XT a shot even thougj i always been a nvidia user.

    Companies have roadmaps and the leak out all the time, you need to plan years in advance though I admit there have been a lot of Nvidia rumours lately with new cards poppin up and dissapearing just as fast again, guess the people stuck at home have nothing better to do atm.

  16. Companies have roadmaps and the leak out all the time, you need to plan years in advance though I admit there have been a lot of Nvidia rumours lately with new cards poppin up and dissapearing just as fast again, guess the people stuck at home have nothing better to do atm.

    When was the last time… please because I really don’t recall. That Nvidia was starting to leak rumors of its next generation cards before even the Ti Variants for the current generation were even launched?

    This is Early by any stretch of the imagination and something that I am surprised to see. It says to me that Nvidia is being reactionary because they have to and it doesn’t play well for them to be reactionary. ESPECIALLY to a company with 1/4th their market cap.

  17. When was the last time… please because I really don’t recall. That Nvidia was starting to leak rumors of its next generation cards before even the Ti Variants for the current generation were even launched?

    This is Early by any stretch of the imagination and something that I am surprised to see. It says to me that Nvidia is being reactionary because they have to and it doesn’t play well for them to be reactionary. ESPECIALLY to a company with 1/4th their market cap.

    and you know nvidia is doing this how?

  18. By the uncharacteristic actions they are taking. Fully admit this is opinion based on only 30 years as a computer enthusiast and IT professional. I could be wrong.

    I expect that they’re making calculated moves, but leaks are a potential problem for every company; there’s no way to discount the possibility that they are or are not intended.
    Worse, the only leak anyone wants to hear is that there’s stock on the way.

  19. By the uncharacteristic actions they are taking. Fully admit this is opinion based on only 30 years as a computer enthusiast and IT professional. I could be wrong.

    So speculation, ok.

    Now speculating myself, I’d say that nvidia hasn’t launched cards on the latest manufacturing process, since… I can’t remember. So this would be a first in quite a while.

    While ampere does come in a 7nm process in its GA100 flavor, its anyone’s guess if they can move the rest of the family. It does seem like the logical choice, but maybe, maybe not.

    Thing is that AFAIK, TSMC 5nm is better than 7nm on yields because of EUV, so there’s a chance that there will be much bigger gains moving to it but nvidia has taken a wait and see approach before moving to a newer node and it worked pretty well up until ampere 8nm samsung node.

Leave a comment