Tuesday, September 20th 2022
NVIDIA Project Beyond GTC Keynote Address: Expect the Expected (RTX 4090)
NVIDIA just kicked off the GTC Autumn 2022 Keynote address that culminates in Project Beyond, the company's launch vehicle for its next-generation GeForce RTX 40-series graphics cards based on the "Ada" architecture. These are expected to nearly double the performance over the present generation, ushering in a new era of photo-real graphics as we inch closer to the metaverse. NVIDIA CEO Jensen Huang is expected to take center-stage to launch these cards.15:00 UTC: The show is on the road.15:00 UTC: AI remains the center focus, including how it plays with gaming.
15:01 UTC: Racer X is a real-time interactive tech demo. Coming soon.
15:02 UTC: Future games will be simulations, not pre-baked- Jensen Huang15:03 UTC: This is seriously good stuff (RacerX). It runs on a single GPU, in real-time, uses RTX Neural Rendering15:05 UTC: Ada Lovelace is a huge GPU15:06 UTC: 76 billion transistors, over 18,000 shaders. 76 billion transistors, Micron GDDR6X memory. Shader execution reordering is major innovation, as big as out-of-order execution for CPUs, gains up to 25% in-game performance. Ada built on TSMC 4 nm, using 4N, a custom process designed in together with NVIDIA.
There's a new streaming multiprocessor design, with a total of 90 TFLOPS. Power efficiency is doubled over Ampere.
Ray Tracing is on the third generation now, with 200 RT TFLOPS and twice the triangle intersection speed.
Deep Learning AI uses 4th gen Tensor Cores, 1400 TFLOPS, "Optical Flow Accelerator"15:07 UTC: Shader Execution Reordering similar to the one we saw with Intel Xe-HPG15:08 UTC: Several new hardware-accelerated ray tracing innovations with 3rd gen RTX.15:09 UTC: DLSS 3 is announced. It brings with it several new innovations, including temporal components, and Reflex latency optimizations. Generates new frames without involving the graphics pipeline.15:11 UTC: Cyberpunk 2077 to get DLSS 3 and SER. 16 times increase in effective performance using DLSS 3 vs. DLSS 1. MS Flight Simulator to get DLSS 3 support15:13 UTC: Portal RTX, a remaster just like Quake II RTX, available from November, created with Omniverse RTX Remix.15:14 UTC: Ada offers a giant leap in total performance. Everything has been increased 40 -> 90 TFLOPS shader, 78 -> 200 TFLOPS RTX, 126 -> 300 TFLOPS OFA, 320 -> 1400 TFLOPS Tensor.15:17 UTC: Power efficiency is more than doubled, but power goes up to 450 W now.15:18 UTC: GeForce RTX 4090 will be available on October 12, priced at $1600. It comes with 24 GB GDDR6X and is 2-4x faster than RTX 3090 Ti.15:18 UTC: RTX 4080 is available in two versions, 16 GB and 12 GB. The 16 GB version starts at $1200, the 12 GB at $900. 2-4x faster than RTX 3080 Ti.15:19 UTC: New pricing for RTX 30-series, "for mainstream gamers", RTX 40-series "for enthusiasts".15:19 UTC: "Ada is a quantum leap for gamers"—improved ray tracing, shader execution reordering, DLSS 3.15:20 UTC: Updates to Omniverse
15:26 UTC: Racer X demo was built by a few dozen artists in just 3 months.15:31 UTC: Digital twins would play a vital sole in product development and lifecycle maintenence.15:31 UTC: Over 150 connectors to Omniverse.15:33 UTC: GDN (graphics delivery network) is the new CDN. Graphics rendering over the Internet will be as big in the future as streaming video is today.15:37 UTC: Omniverse Cloud, a planetary-scale GDN15:37 UTC: THOR SuperChip for automotive applications.15:41 UTC: NVIDIA next-generation Drive
15:01 UTC: Racer X is a real-time interactive tech demo. Coming soon.
15:02 UTC: Future games will be simulations, not pre-baked- Jensen Huang15:03 UTC: This is seriously good stuff (RacerX). It runs on a single GPU, in real-time, uses RTX Neural Rendering15:05 UTC: Ada Lovelace is a huge GPU15:06 UTC: 76 billion transistors, over 18,000 shaders. 76 billion transistors, Micron GDDR6X memory. Shader execution reordering is major innovation, as big as out-of-order execution for CPUs, gains up to 25% in-game performance. Ada built on TSMC 4 nm, using 4N, a custom process designed in together with NVIDIA.
There's a new streaming multiprocessor design, with a total of 90 TFLOPS. Power efficiency is doubled over Ampere.
Ray Tracing is on the third generation now, with 200 RT TFLOPS and twice the triangle intersection speed.
Deep Learning AI uses 4th gen Tensor Cores, 1400 TFLOPS, "Optical Flow Accelerator"15:07 UTC: Shader Execution Reordering similar to the one we saw with Intel Xe-HPG15:08 UTC: Several new hardware-accelerated ray tracing innovations with 3rd gen RTX.15:09 UTC: DLSS 3 is announced. It brings with it several new innovations, including temporal components, and Reflex latency optimizations. Generates new frames without involving the graphics pipeline.15:11 UTC: Cyberpunk 2077 to get DLSS 3 and SER. 16 times increase in effective performance using DLSS 3 vs. DLSS 1. MS Flight Simulator to get DLSS 3 support15:13 UTC: Portal RTX, a remaster just like Quake II RTX, available from November, created with Omniverse RTX Remix.15:14 UTC: Ada offers a giant leap in total performance. Everything has been increased 40 -> 90 TFLOPS shader, 78 -> 200 TFLOPS RTX, 126 -> 300 TFLOPS OFA, 320 -> 1400 TFLOPS Tensor.15:17 UTC: Power efficiency is more than doubled, but power goes up to 450 W now.15:18 UTC: GeForce RTX 4090 will be available on October 12, priced at $1600. It comes with 24 GB GDDR6X and is 2-4x faster than RTX 3090 Ti.15:18 UTC: RTX 4080 is available in two versions, 16 GB and 12 GB. The 16 GB version starts at $1200, the 12 GB at $900. 2-4x faster than RTX 3080 Ti.15:19 UTC: New pricing for RTX 30-series, "for mainstream gamers", RTX 40-series "for enthusiasts".15:19 UTC: "Ada is a quantum leap for gamers"—improved ray tracing, shader execution reordering, DLSS 3.15:20 UTC: Updates to Omniverse
15:26 UTC: Racer X demo was built by a few dozen artists in just 3 months.15:31 UTC: Digital twins would play a vital sole in product development and lifecycle maintenence.15:31 UTC: Over 150 connectors to Omniverse.15:33 UTC: GDN (graphics delivery network) is the new CDN. Graphics rendering over the Internet will be as big in the future as streaming video is today.15:37 UTC: Omniverse Cloud, a planetary-scale GDN15:37 UTC: THOR SuperChip for automotive applications.15:41 UTC: NVIDIA next-generation Drive
333 Comments on NVIDIA Project Beyond GTC Keynote Address: Expect the Expected (RTX 4090)
Outside of this being a marketing/branding move that borders very close on being exploitative in and of itself, this has other problems: the Titan branding also brought with it price separation. Titans were wildly expensive compared to their close Geforce siblings, which was defensible through these being for business, not for gaming. Through including these same cards into the Geforce lineup, they now have the "freedom" to lift pricing for the rest of the Geforce lineup up so that instead of being a major step upwards in price, it's instead continuous with lower tier cards. That's how you go from $1200 Titan X(p) and $700 1080 Ti (+71% price) to $1600 RTX 4090 and $1200 4080 16GB (+33%). The net effect of this is not the Titan-class card being "a lot cheaper than the Quadro series" - they always were - but instead the entirety of the Geforce lineup becoming more expensive through a persistent lifting of the price ceiling for such cards, and thus slowly shifting the marketing/pricing equivalent of the Overton window - the window of what is seen as acceptable and reasonable pricing for a GPU.
There is no other explanation of this that makes any type of sense other than Nvidia wanting to increase prices and squeeze more money out of gamers.
So by this logic, if we should pay 700-800USD to an RTX 3060 we should be happy, by a 400usd "discount" as it should be? I do not think so....
Only moderate price increase acceptable, that covers inflation, design and manufacturing complexity. But by the age of mining and high margin of nVidia, the last years of price increasing is not acceptable.
Just simply look at the financial balance of these companies nVidia, AMD... Can you realize a similar financial increasement in your personal live, or in your familly or friends? They are growing as hell while an average person in the society have even less financial possibilities than years before.
What does this mean? That GPUs are getting increasingly unaffordable, and the main culprit is that GPU makers simply aren't making and selling affordable GPUs, instead insisting on artificially inflating margins. The RTX 3050 could easily be $150 if Nvidia was focused on designing an affordable GPU. These are conscious product segmentation choices made on the level of chip design and tiering. Instead, Nvidia is working concertedly towards increasing prices across the board. And sadly, so far AMD is happy to follow suit.
I don't care what others can or can't afford if they're too expensive for you or anyone else don't buy them it's simple these aren't necessities in life we are talking about.
Same with me if the 80/90 tier ever goes out of my price range guess what I'll stop buying them life goes on...
Getting upset over how a company prices their own products is pointless... They're not a charity or your friends they will all try to make the most money possible.
I think we all know how both of these turned out.
NVIDIA can do whatever they want. But I will listen to this advice and simply do not buy it. If they drop prices at some point, I might. If not, I will wait for the next generation (or at least consider a Radeon card, although I do want great RT performance and I do love DLSS).
@kiakk Calling me out about a flagship being cheaper than a previous gen flagship and then going on a tangent about 60 tier pricing and how it should only be priced in a way that is ok with him makes no sense to me though. But what other choice do we have? I mean if you don't like the price the only option is don't buy it. Getting upset in a tech forum about it does absolutely nothing.
The 4080 12GB is stupidly priced at 900 usd I wouldn't touch it and I doubt reviews will change that.. That leaves me with the 4080 16gb and 4090 if I want a noticeable upgrade this gen... Will I buy one no idea but I will at least reserve judgment on them till 3rd party reviews go live. I still think this has more to do with them wanting to clear out ampere stock I mean they still consider 3 of them part of their current offerings as it is.
I personally like RT especially in games like Control and Cyberpunk with 3-4 different implementation so AMD is probably not an option this generation again. Sounds like their implementation will be quite a bit weaker but I'd love to be wrong.
I mean, there are other choices. Used market, consoles (a gift to AMD, really), finding other hobbies. Nvidia has nothing to gain from alienating customers - but that's what they're doing. They've been pushing hard for quite a while to get as much money as they can from a rapidly expanding customer base, but we've already reached a saturation point, and rather than accepting that they overplayed their hand, or played hard and went out of bounds, and accepting the (small!) cost of this, Nvidia are instead continuing on a path of maximizing profits rather than designing attractive products people can actually afford.
A lot of Ada design was probably decided during the peak of the mining boom I wonder in retrospect if they would change anything I'm sure they are still salty with retailers making more than they did the first 6-8 months of amperes life. I personally expected the two 4080s to be cheaper but the 4090 to be more expensive but here we are.
Going forward I'm as interested as anyone to how the market reacts Turing didn't do so well but ampere did amazing for them due to reasons. It'll be interesting how Ada does.
Kudos to those who got a 10GB 3080 near msrp at launch though in retrospect that was a hell of a deal same with the 6800XT to a lesser extent.
To me, this is part of why I tend towards believing EVGA's version of their split: their description is consistent with Nvidia's behaviour towards only caring about increasing margins, ASPs and per-tier pricing for more than half a decade now. It's likely not the full story, but that glove definitely does seem to fit nonetheless.
Of course, mobile also likely plays a big part in this - and the relatively large dice and expensive designs for the past few generations match well with these being targeted towards lower clocked mobile chips that still deliver good absolute performance. Smaller, cheaper designs would inevitably do worse in mobile by simple virtue of not having the ability to clock as high. And, of course, two successive crypto booms and a two-year gloabl lockdown have no doubt also driven a belief that whatever the price, the GPUs will sell. We'll see what kind of correction this leads to in the near future, if anything. For now, their expressed tactic is to not change anything and try to rake in profits as much as possible. Which really does leave the door open for AMD.
I have worked extremely hard my entire adult life(I am now 44) to be in the position where my hobbies aren't a financial hardship, and this is just one example where that work has paid off.
Also, 6-tier GPUs are distinctly not Ferraris - they're supposed to be the Toyotas of the GPU world. And if supposedly cheap Toyotas start being priced like Ferraris, there's something serious wrong going on.
Still, I think Nvidia (and AMD for that matter, though they seem a tad more flexible currently) needs to shift things. If they just made the leap, told people that "what used to be called 6 is now called 5", and accepted the dampening effect of this on sales for one generation (which they could most likely counteract by marketing the crap out of that new 50-series), they'd have a lot more flexibility in product segmentation. Instead they're forcing themselves to fit a massive range of performance into just four numbered tiers. Which is just stupid. So instead we've now got ... what, eight SKUs across four tiers, ranging from the 3060 to 3090 Ti. The 3050 is clearly an afterthought, and a card they don't really seem to want to sell, given its pricing and market positioning.
As I suggested above, they could also have alleviated this by creating a "premium" brand of some sort, moving perhaps the top three SKUs to this tier. Call them Titan, call them GeforceX, call them Xtreme XGeforce XRTX - whatever. This way they could have called the 3050 the 3060, priced it at $200, and sold tons and tons of them. And they could still have had $700 80-tier premium/high end tier cards that would be aspirational for the people buying $200 GPUs - most of those spend far less on their PC than the cost of a 3090, so selling them on the 3090 being great isn't much of an advertisement anyhow.
Of course, the mining+lockdown WFH/gaming booms have made such shifts unnecessary - up until now. I guess we'll see how this plays out, but if they insist on keeping the "Geforce [four numbers]" structure as their only GPU brand, they need to start shifting things - the sooner the better for them. But for now, their preferred tactic seems to be "profits today, screw tomorrow".
But maybe it is time to change things. If cards are too expensive to manufacture, then they should make smaller GPUs with simpler PCBs and coolers. Maybe we should go back to 200 W being the maximum power consumption. Make smaller performance increases, but make it possible for people to actually buy cards in the range of $200 to $500.
All other PC components are super affordable, including CPUs, which have had gigantic leaps in performance in the last 5 years, but prices have basically stayed the same. So it can be done.
I also agree that this situation presents a massive opportunity for AMD. Regardless of where RDNA3 peaks in terms of absolute performance, if they flesh out the $300-700 segment with high performing, good value options relatively quickly, they have an unprecedented opportunity to overtake Nvidia - as this pitch from Nvidia (as well as their earnings call) confirms that they've got more Ampere stock than they know what to do with, and they don't want to stomach the cost of cutting prices unless they really have to. If RDNA3 actually delivers its promised 50% perf/W increase, and die sizes and MCM packaging costs allow for somewhat competitive pricing, this could be the generation where AMD makes a significant move up from their perennial ~20% market share. That's my hope - but knowing AMD and their recent preference for high ASPs and following in Nvidia's pricing footsteps, I'm not that hopeful. Fingers crossed, though. This is absolutely true. There's also the simple fact of GPUs largely being good enough for quite a lot of things for quite a while. We're no longer seeing anyone need a GPU upgrade after even 3 years - unless you're an incorrigible snob, at least. Lots of factors play into this, from less growth in graphical requirements to the massive presence of AA/lower budget/indie games in the current market to open source upscaling tech to lower resolutions in many cases still being fine in terms of graphical fidelity. It could absolutely be that Nvidia is preparing for a future of (much) lower sales, but I don't think they can do that by pushing for $1500 GPUs becoming even remotely normal - the number of people affording those isn't going to increase anytime soon. Diversifying into other markets is smart, though, and AMD also needs to do the same (not that they aren't already). But the impending/already happened death of the 1-2 year GPU upgrade cycle won't be solved by higher prices and more premium products.
It was my logic, nVidia is using PR to get gamers to buy the 40x0 products too. And successful. The explanation lies within the gamer community. Professionals mostly knew, what they buy and what for. I agree partially. RT on gaming cards was an innovation, based on technology, nVidia used in the professional markets. They implemented it in gamer hardware to realize synergies and market leadership (AMD is clearly behind in this area). I remember all the discussions about the 20x0 series. My position was: wait some years, and there will be no high end card without ray tracing. This guarantees nVidia technological leadership in the gaming sector without additional development costs, as they need rt/cuda/tensor for the hpc sector anyway. So what is going on? nVidia raises the limits between semi-professional use and high-end gaming, to serve different peer groups with one product. And this from the beginning. Cudacore support was even enabled in the gaming drivers back to maxwell. Every student, who needs raytracing or cudacore power could do some crazy calculations with cheap gaming cards. My first nVidia card was a 970 for exactly that reason, therefore I had radeon. Once you are adapted to a special software/hardware environment, you will not change easily.
People who spend >1000$ for gaming, well, its ok. What I dont like is nVidias low and midtear pricing. But that's an opportunity for AMD. So there is something for everyone.
That being said I still expect the 7900XT to be 1200 usd or less and the 7800XT to be 800 usd or less I just don't think AMD cares as long as its 50-100 usd cheaper than a comparable Nvidia sku even if its much cheaper to make. I also wouldn't be surprised if the 7900XT was 1500 usd though..... I doubt anything lower than these 2 cards will launch anytime soon though with maybe a 7700XT early next year that to me will be the more interesting card in how they price it.
Another thing and this goes for Nvidia as well is TSMC is really the one dictating prices. Not sure if true but the New Ada gpus are supposedly much more costly to make than Ampere I really would like to see a breakdown of component cost for sure though I know SMD are getting more expensive as well.