Turing Discussion Thread

[SteelCrysis]

RTX 2070 coming on October 17 for $499: https://www.techpowerup.com/247933/nvidi...tober-17th

[SteelCrysis]

RTX 2080 Ti delayed again: https://www.techpowerup.com/247962/nvidi...er-5th-9th

[SteelCrysis]

https://www.pugetsystems.com/blog/2018/0...lows-1233/

What NVIDIA is presenting with the RT cores in their new Turing GPU architecture is a way to achieve much more natural results natively. With this new hardware, they claim the above scenarios will “just work” in real time. We still need to see how this plays out with final hardware and software updates, but the potential benefits are enormous.

The video game industry is already jumping into this tech, but since they will still need to support older, non-RTX hardware for quite a while, they will still need to use their current workflows until RTX enabled hardware becomes widespread. However, as this technology gains support in professional applications, someone working in architectural visualization could potentially use it to go from their blueprints to final render much faster with more lifelike results. Since this would all work in real time, they can be sitting with their client and change materials, add furniture, move walls, etc - and everything will update live. A studio producing a weekly animated show might be able to render the entire episode in real time, allowing them to spend more time in production. Even high end movie studios, which will likely still rely on traditional raytraced renderers, could have a near final graphics “preview” with the ability to make significant visual changes and not have to wait to re-render everything.

This only scratches the surface of what creative studios will accomplish with these cards. We have yet to see how the new RT Cores will improve traditional raytraced rendering. Or what the scientific communities will do with Tensor cores, which are making their mainstream debut in the GeForce RTX series. The potential presented by these cards reaches far beyond the benchmarks of old-school rasterized game technology which most people are focused on.

[SteelCrysis]

https://www.techpowerup.com/247985/nvidi...ood-enough

As you can see, single-monitor idle power consumption is much better now, bringing it to acceptable levels. Even though the numbers are not as low as Pascal, the improvement is great, reaching values similar to AMD's Vega. Blu-ray power is improved a little bit. Multi-monitor power consumption, which was really terrible, hasn't seen any improvements at all. This could turn into a deal breaker for many semi-professional users looking at using Turing, not just for gaming, but productivity with multiple monitors. An extra power draw of more than 40 W over Pascal will quickly add up into real Dollars for PCs that run all day, even though they're not used for gaming most of the time.
...
The table above shows monitored clocks and voltages for the non-gaming power states and it looks like NVIDIA did several things: First, the memory frequency in single-monitor idle and Blu-ray has been reduced by 50%, which definitely helps with power. Second, for the GTX 2080, the idle voltages have also been lowered slightly, to bring them in line with the idle voltages of RTX 2080 Ti. I'm sure there's additional under-the-hood improvements to power management, internal ones, that are not visible to any monitoring.

Let's just hope that multi-monitor idle power gets addressed soon, too.

[SteelCrysis]

https://www.tomshardware.com/news/amd-ra...37848.html

The arrival of Nvidia’s GeForce RTX graphics cards has left many consumers underwhelmed, with the promise of a ray tracing-enabled tomorrow not doing much for enthusiast PC gamers today. With ongoing rumblings about new AMD graphics cards on the horizon, many are awaiting the Red Team’s response to Team Green’s next-gen RTX graphics cards.

There are many reasons AMD could make waves in the GPU market next year, which we'll discuss below, but most of the following points are hypothetical or opinion (based on what we know today). A lot depends on how things could go for AMD at this point. We of course don't know exactly how things willplay out. But let's consider what 2019 and beyond might look like for AMD on the graphics card front.
...
With Nvidia’s focus on ray tracing, its high RTX-card price points, and less-than-stellar generational gains for traditional rasterized gaming performance, AMD has been handed an opportunity to close the gap or even take the lead over its competitor in the GPU market. As we’ve noted, this will all depend on where the aforementioned chips land in the performance and price stack when AMD launches its next generation graphics architecture. The only thing we know for sure at this point is that AMD has new graphics cards coming (likely in the first half of 2019). While we'd still put AMD in underdog territory on the graphics side, its recent successes on the CPU front with Ryzen proves the company can deliver an exciting, competitive product after years of less-than-stellar silicon.

https://www.techpowerup.com/247958/nvidi...ange-users

t won't be easy to see an AMD counterpart to NVIDIA GeForce RTX 2080 Ti in the short time. Heck, it will be difficult to see any real competitors to the RTX 2000 Series anytime soon. But maybe AMD doesn't need to compete on that front. Not yet. The reason is sound and clear: RTX prices are NVIDIA's weakness, and that weakness, my fellow readers, could become AMD's biggest strength.
...
That won't probably be enough to make the hypothetical RX 680 catch up with a GTX 1070: performance of the latter is +34% the one we found in the RX 580 on average according to our tests, so even with that refresh we will have a more competitive Radeon RX family that could win the price/performance battle, and that is no small feat.

The new cards would also not target just existing GTX 7/9 Series users, but also those older AMD Radeon users that were expecting a nice upgrade on performance without having to sell their souls. And for the undecided users, the ones that are thinking about getting a GTX 1050/Ti or a GTX 1060, AMD's offer could be quite attractive if price/performance ratio hits NVIDIA where it hurts more.

That would put that new family of graphic cards (Radeon RX 600?) on a pretty good position to compete with GeForce GTX 1000. NVIDIA presumably could still be king in the power consumption area, but besides that, AMD could position itself on that $300-$500 range (and even below that) with a really compelling suite of products.

So yes, AMD could have a winning hand here. Your move, AMD.

[SteelCrysis]

https://www.extremetech.com/gaming/27845...h-a-turing
Be sure to read the whole article, it really says it all about Turing.

If the RTX 2080 had come in at GeForce 1080 pricing and the RTX 2080 Ti had slapped $100 – $150 on the GTX 1080 Ti, I still wouldn’t be telling anyone to buy these cards expecting to dance the ray-traced mamba across the proverbial dance floor for the next decade. But there would at least be a weak argument for some real-world performance gains at improved performance-per-dollar ratios and a little next-gen cherry on top. With Nvidia’s price increases factored into the equation, I can’t recommend spending top dollar to buy silicon that will almost certainly be replaced by better-performing cards at lower prices and lower power consumption within the next 12-18 months. Turing is the weakest generation-on-generation upgrade that Nvidia has ever shipped once price increases are taken into account. The historical record offers no evidence to believe anything below the RTX 2080 Ti will be a credible performer in ray-traced workloads over the long term.

ExtremeTech, therefore, does not recommend purchasing either the RTX 2080 or RTX 2080 Ti for their hoped-for performance in future games. Ray tracing may be the future of gaming, but that future isn’t here yet. Mainstream support for these features will not arrive on 12nm GPUs and will not be adopted widely in the mass market before the next generation of cards. And again — every dollar you drop on $700 and $1,200 GPUs is a dollar that tells Nvidia it ought to be charging you more for your next graphics card.

The RTX 2080 and 2080 Ti are the beginning of a new approach to rendering. Long-term, the technological approach they champion may be successful. But that doesn’t make either of them a good buy today, while the historical record suggests no ground-breaking GPU is a particularly good buy its first time out of the gate. Ironically, that’s entirely Nvidia’s fault. If the company was less-good at immediately overtaking its own previous GPU architectures will new hardware that offers dramatically better performance and price/performance ratios, I’d have less of a reason to take the stance I’m taking. But between the price increases and Nvidia’s own historical performance, the smart move here is to wait. Whether you’re waiting for sane pricing, 7nm, or just happy with your current GPU — that’s up to you.

[SteelCrysis]

https://www.techpowerup.com/248382/msi-t...-lightning

Back on September 27th, MSI talked candidly with PConline at the MSI Gaming New Appreciation Conference, in Shanghai. Multiple MSI executives were available to answer questions regarding products, launches, and potential issues. The first question asked was about the brewing US-Chinese trade war and if it will affect prices of graphics cards and CPUs. To which, Liao Wei, Deputy General Manager of MSI Global Multimedia Business Unit, and MSI Headquarters Graphics Card Products gave an actual answer. Stating that the since NVIDIA's GPU core is handled by a TSMC in Taiwan and memory is handled by Samsung and Hynix in South Korea and the United States respectively, there is little chance of further graphics card price hikes. However CPU side prices may increase on the Intel side, however, AMD is expected to be unaffected.

They were next asked why NVIDIA's GeForce RTX graphics cards are already out of stock and hard to come by, To which Liao Wei again replied, stating that the RTX 2080 shipments are actually holding quite well, however, the RTX 2080 Ti shipments have been relatively small which is due to production yields. While this is to be expected considering the large size of the GPU die it also means NVIDIA appears to be focusing more on RTX 2080 production. Liao Wei also touched on the fact that the RTX series is far more complex than the previous generation, with the RTX 2080 Ti using some 2600 components compared to the GTX 1080 Ti at 1600. For further comparison, it is said the RTX 2080 uses 2400 components and the RTX 2070 some 2200. So not only are the newer GPUs more complex and made of more components, but they also face a production crunch. Originally it was possible to produce 3500 GPUs a day for the GTX 1080 Ti, while the RTX 2080 Ti can only be produced at a rate of 1800 per day. It is likely this is before screening for defective chips.

[SteelCrysis]

https://www.tomshardware.com/reviews/nvi...51-11.html

According to Nvidia, GeForce RTX 2070 partner boards should be available at $500 as well. How likely is that to happen? Well, GeForce RTX 2080 was supposed to start at $700, and today most third-party cards come closer to $800. We were told GeForce RTX 2080 Ti would debut at $1000. But those are all listed for $1200 or more, and universally out of stock. Further, multiple partners were supposed to send GeForce RTX 2070 samples for today’s launch, and not one was able to deliver. Pardon our skepticism, but we’re drawing from the previous rocky launch, current pricing of month-old Turing-based cards, and the inability of Nvidia’s partners to ship their own designs ahead of claimed retail availability.

It’s true that GeForce RTX 2070 Founders Edition is faster than GeForce GTX 1080 across the 13 games and two resolutions we tested, sometimes by double-digit percentages. The TU106-based card also beats Radeon RX Vega 64 in most performance benchmarks. Although AMD’s flagship does bag a few wins across our suite, any attempt to compare power consumption or efficiency penalizes the Radeon significantly. At $500, then, the GeForce RTX 2070 does make sense as a replacement for GeForce GTX 1080, even if its value proposition is only slightly better.

We’re not reviewing a $500 card, though. The GeForce RTX 2070 Founders Edition sells for $600, which is $100+ higher than the cheapest GTX 1080s and $50 more than Nvidia’s own GeForce GTX 1080 Founders Edition. It’s difficult to get excited about RTX 2070 at $600, unfortunately.

Turing introduces a lot of novel functionality primed to improve the realism of gaming in the weeks and months to come. So, what’s up with the lackluster reception of Turing-based graphics cards? It’s a three-part interplay of hyped-up technology that can’t be used yet, comparisons to plentiful Pascal-based cards, and a resulting (negative) perception of value. In generations past, Nvidia gave us more performance at a comparable, if not better price. This time around, the company is mostly competing against its own cards with MSRPs reflecting a lack of competition. GeForce RTX 2070 is basically a step sideways for anyone who was previously eyeballing GeForce GTX 1080. GeForce RTX 2080 is a step sideways for anyone with a GeForce GTX 1080 Ti. Only GeForce RTX 2080 Ti sets itself apart as an unrivaled winner among folks who were previously willing to pay $1200 for Titan-class frame rates. Buy Pascal or buy Turing; Nvidia wins either way.

And that’s why, so far, GeForce RTX 2000-series cards aren’t generating the same levels of enthusiasm with gamers as they are with game developers. Nvidia seems to be counting on the first round of ray tracing- and DLSS-enabled games to validate its architectural decisions. Based on our hands-on experiences with various demos, we have no doubt that anyone with a GeForce RTX 2080 Ti will enjoy their purchase. But it’s not clear yet whether RTX 2070 is fast enough to make Turing’s most newsworthy features usable. At this point, we’d hold off on a purchase unless you were already in the market for a GeForce GTX 1080/Radeon RX Vega 64-class card and can find GeForce RTX 2070 priced competitively.

It’s a foregone conclusion that we’ll have to revisit our opinions of all three GeForce RTX 2000-series cards once game developers start rolling out their Turing-optimized titles. Will ray tracing make enough of a difference to compel a graphics upgrade? Does GeForce RTX 2070 have enough RT cores to maintain playable performance in DXR-enabled games? Will broader availability force today’s inflated prices down? We sure hope so. Nvidia isn’t making any more Pascal GPUs, so the good deals on previous-gen cards are bound to run out sooner than later.

[SteelCrysis]

https://www.extremetech.com/gaming/27893...e-increase

Ostensibly, the 2070 should be a $500 part, but reviewers are flatly telling readers not to expect any cards in-market for that price any time soon. Instead, these GPUs will launch at $600 and above, which means they’ve effectively left the GTX 1080 behind and are bidding to challenge the GTX 1080 Ti. And as of today, you can buy a GTX 1080 for $434 on Newegg, which means the 2070 needs to demonstrate some significant performance gains to match it. With a base price of $600, we’re looking for the RTX 2070 to offer at least 1.38x more performance than the GTX 1080 to be a decent buy.

The chart — not to mention reviews — confirm this isn’t something the RTX 2070 can actually do. Anandtech delivers the nasty news in a single paragraph.
...
The comparison doesn’t get any better if we check relative generation-to-generation upgrades for the x70 family. Anand points out that while the GTX 1070 delivered a 1.57x increase in 1440p performance relative to the GTX 970, the GTX 2070 can only manage a 1.36x increase over the GTX 1070. Now, watch the math. We’ll use the higher-priced FE cards to make the comparison look better for Nvidia.

GTX 970 launch price: $329
GTX 1070 launch price: $449
GTX 2070 launch price: $600

In 2016, you could buy a 1.57x increase in performance for a 1.36x increase in price. Today, you can buy a 1.36x increase in performance for a 1.33x increase in price relative to the launch price of the GTX 1070. Except, of course, the GTX 1070 isn’t a launch card anymore. It currently sells for $339 — which means the actual price/performance ratio for this comparison is a 1.76x increase in price for a 1.36x increase in performance.

This is not compelling. It will never be compelling.
...
But even this generous assumption leaves the RTX 2070 falling flat. It has just 78 percent the ray tracing performance of the RTX 2080, which means gamers can look forward to bumping along at ~47fps in 1080p with their brand-new $600 video card. As I discuss in our review of the RTX 2080 and RTX 2080 Ti, history overwhelmingly indicates that taking a bet on a next-gen feature by trying to buy in at a low level rarely works. If you don’t believe me, ask an 8800 GTS owner if they were happy to see the GPUs they bought for $400 in 2006 blown out of the water by a $300 card barely a year later.
...
I’ve made it clear that I don’t think gamers should buy Turing GPUs. The price increases aren’t remotely justified by available performance and the future value of ray tracing and DLSS is a complete unknown. But I will say this: The more you care about ray tracing and performance in that mode, the leerier you should be of buying an RTX 2070. If you doubt this, do yourself a favor, and go look at reviews of early DirectX 10 and even some DX11 GPUs, like the HD 5770. It’s common for midrange cards to struggle in new APIs. And ray tracing, as Nvidia has noted, is far more than just an API change — it requires extensive silicon support, specialized hardware, and capabilities that are only now in their infancy.

If you really care about future performance, wait. Don’t buy the sizzle. Don’t buy the idea that Nvidia wouldn’t ship a GPU with technical ray tracing ability that couldn’t practically be used for ray tracing. It’s entirely possible that games will offer different levels of ray tracing to target different GPUs, but since we don’t know if that’ll happen yet or how good it will look if it does. The RTX 2070 might end up delivering rather uninspiring support levels while the jaw-dropping effects in Battlefield V require a GPU $200 – $600 more expensive.

https://www.hardocp.com/article/2018/10/..._review/12

Update: MSI has contacted us today asking HardOCP to remove this RTX 2070 review, even though MSI had nothing to do with sourcing this review. NVIDIA's green feathers are apparently flying over this RTX 2070 review being published before its embargo date and time that is has with reviewers that signed its NDA. This is how things turn out when NVIDIA tries to force 5 year blanket NDAs down journalists throats. We chose not to sign NVIDIA's NDA. Our review is 100% legitimate and we are not going to remove it because NVIDIA is throwing a fit over it being published. The fact of the matter is that NVIDIA changed its entire NDA/Product Embargo structure after we reported on GPP this year. It did this to muzzle stories about NVIDIA in the future, and it is on NVIDIA for tying that to its product reviews. It is sad that MSI is having to deal with the brunt of NVIDIA's fury over this, and to that, we are sorry that is happening. This review could have easily been over any other AIB's card, it just so happens that an MSI card was the first one that we could source.

[SteelCrysis]

https://www.tomshardware.com/reviews/nvi...,5855.html

But here we are more than two months out from the Turing launch, and Nvidia doesn’t have a whole lot to show for its effort, other than a trio of high-priced cards that are consistently failing to arrive in stores at their promised starting prices.
...
A quick look at Newegg and Amazon tells us that most RTX 2080 cards are selling for more than the premium Founders Edition pricing (north of $800 in the US and £600 in the UK), nowhere near $700. And the same is generally true of third-party RTX 2080 Ti cards—if you can find them in stock. As we wrote this on the official launch day of the RTX 2070, third-party variants of that card weren’t yet showing up for sale. And Nvidia’s own product page had a “Notify Me” button for its $599 FE variant (£549 UK), which will presumably arrive on October 17th as promised. But recent history tells us it’s extremely unlikely that we’ll see a 2070 card anywhere near its promised $499 starting price anytime soon.

Of course, card scarcity and inflated prices are nothing new for graphics card launches—I remember paying about $30 (22.74 pounds) over MSRP for my ATI Radeon HD 2600 XT in the weeks after it launched in mid-2007. But a $30 premium is a lot easier to tolerate than $100 or more.
...
The bigger issue is why Nvidia is, in effect, over-promising and under-delivering with its RTX lineup--at least here in the early days of Turing card availability. Of course the company will rightly say it doesn’t have control over the price of third-party cards or when developers push out new games or updates. But with no new high-end cards from AMD on the market yet either, it’s unclear why Nvidia launched these cards when it did, rather than wait a few months. And of course it could have sold its own Founders Edition cards at slightly lower prices, pushing its third-party partners to be competitive.

We’re sure to see more than a few enticing games with ray tracing and DLSS features by 2019. By then the stock of 10 series cards may have dried up, and we’ll be able to actually evaluate the RTX cards on their key new features.

Until then, many looking for value in the high-end card market will likely keep looking to older cards--or waiting to see what AMD can deliver with Navi. Even if Navi only brings enough competition to bring down prices on today’s Turing offerings, that will be a boon for Nvidia’s sales numbers. Hopefully more RTX cards in consumers hands will also entice more game developers to push through features that take advantage of the uniqueness of Nvidia’s Turing hardware.

[SteelCrysis]

https://www.techpowerup.com/248636/nvidi...tx-2070-ti

That said, NVIDIA is not on solid-ground with the RTX 2070, and there's a vast price gap between the RTX 2070 and the $800 RTX 2080. GIGABYTE all but confirmed the existence of an SKU in between.

Called the GeForce RTX 2070 Ti, this SKU will likely be carved out of the TU104 silicon, since NVIDIA has already maxed out the TU106 with the RTX 2070. The TU104 features 48 streaming multiprocessors in comparison to the 36 featured by TU106. NVIDIA has the opportunity to carve out the RTX 2070 Ti by cutting down SM count of the TU104 to somewhere in the middle of those of the RTX 2070 and the RTX 2080, while leaving the memory subsystem untouched. With these, it could come up with an SKU that's sufficiently faster than the GTX 1070 Ti and GTX 1080 from the previous generation, and rake in sales around the $600-650 mark.

[SteelCrysis]

https://www.techpowerup.com/248649/remed...erformance

Remedy may be able to reduce that impact in the final version of its engine and in the game, but those 9.2 ms will clearly influence the framerate we can achieve. Playing at 30 fps requires 33 ms and playing at 60 fps requires 17 ms per frame. If we enable NVIDIA's RTX effects that would translate to a framerate of about 40 fps during the game with a 1920x1080 resolution on a GeForce RTX 2080 Ti. The result is excellent visually: clearer shadows and reflections that are independent of the camera and angle show up and give a photorealistic finish to the game, but the cost is high. Too much, maybe?

[SteelCrysis]

https://www.extremetech.com/gaming/27905...ngine-demo

There are some caveats to keep in mind. This is a demo, not a shipping title, and the demo was designed to show off Nvidia’s ray tracing technology, not for maximum performance. But the fact that it opens with an undersampled shimmering scene to start with also suggests that this represented some kind of compromise — if Remedy had been able to remove the effect without harming performance further they surely would’ve done so, if only to make the demo that much more impressive. Meanwhile, Battlefield V seems to demonstrate that at least the RTX 2080 Ti is capable of integrating impressive ray traced visual effects without tanking the frame rate, even if 60ish fps at 1080p isn’t much to write home about by conventional rasterization standards. Clearly, developers have at least some flexibility to tailor their RTX implementations to the frame rates and experiences they wish to target.

At the same time, however, the little evidence we do have is all stacking up on the wrong side of the equation. Nvidia is the only player in this game that could clear up the confusion, but the company has released no performance estimates or clarifying remarks. We don’t know how much fine-grained control users will have to apply ray tracing selectively or even which GPU to recommend to customers who want to take advantage of it.

If you’re wondering why many in the tech press are so lackluster on Turing, it’s not just because Nvidia launched a GPU when games that take advantage of its features haven’t been released. It’s also because the company has denied the press and public any opportunity to objectively evaluate the value of these features. It’s been two months since the RTX family debuted. It is not unreasonable, at this point, for Nvidia to issue a clarifying statement along the lines of: “We expect the RTX 2070 to be capable of delivering the next-generation visual effects our customers expect while maintaining smooth, playable frame rates and will work with developers to ensure all RTX customers can take advantage of these features.” That doesn’t lock the company in to a specific frame rate. It doesn’t guarantee 60fps. It just says “You’re going to get to use the features we’re asking you to pay at least $500 for.”

Nvidia has asked customers to accept significant price increases because, to hear the company tell it, the RTX 2070, 2080, and 2080 Ti will deliver ray traced gaming experiences that justify the premium. Under the circumstances, it’s reasonable to ask the company to demonstrate or pledge that the RTX 2070 will be capable of both ray tracing and reasonable frame rates.

[SteelCrysis]

https://www.techpowerup.com/248769/patch...ustom-bios

BIOS modder Vipeax has released a special patched version of NVFlash (version 5.527.0), the utility that allows you to extract and flash the video BIOS of your NVIDIA GeForce graphics card.
...
PC enthusiasts look to flash their Founders Edition cards with BIOS ROMs of custom-design graphics cards by other NVIDIA add-in card partners, mainly to increase power limits that allow the GPU to sustain boost frequencies better, and increase overclocking headroom. As an obligatory word of caution, use of NVFlash isn't covered by product warranties, and you use it at your own risk, especially when cross-flashing between cards that might have subtle differences. We manually checked the modified executable (not just Virustotal) and it doesn't contain any malware.

https://www.extremetech.com/gaming/27950...-price-tag

But there have consistently been RTX 2070 GPUs in-market for the past week, and they’ve been selling for the $500 price tag. While we will absolutely continue to revisit the question of GPU availability, a week of $500 availability requires that we revisit our own original evaluation and acknowledge that our $600 estimated launch price tag was wrong.
...
Does this make the RTX 2070 a better value than I previously believed it to be? Absolutely, yes. Does it make it a compelling GPU purchase? No — though it may be a reasonable one now, depending on your requirements (more on this in a moment).
...
After reconsidering the RTX 2070 at the appropriate price point, our guidance is this: If you can afford the cost increase and do not care whether the RTX features pan out, the RTX 2070 will deliver a performance improvement over the 1070 that is mostly in-line with what you’ll be asked to pay for it.

If you actually care about ray tracing, can afford a $500 GPU, and wish to make certain that you purchase a card that can deliver this feature, we do not recommend the RTX 2070, for the sole reason that Nvidia has not yet demonstrated that it will. Once Nvidia has made that demonstration if the card’s price and features are attractive to you, pull the trigger.

If you absolutely in need of a new GPU and cannot afford a $500 card, we recommend either buying a used card (for short-term purposes while waiting to see if NV has 7nm GPUs coming in 2019 and/or what Turing RTX performance will actually look like) or buying the least expensive new card you can afford at a reasonable performance level and banking the rest towards a higher-end GPU once these questions are answered. And keep in mind, Pascal and AMD’s Polaris are both still capable of providing solid midrange performance, even if the 10-series and Vega no longer represent the latest and greatest in GPU technology.

[SteelCrysis]

https://www.tomshardware.com/reviews/dls...870-5.html

Inner workings aside, DLSS remains one of the Turing architecture's most interesting capabilities, and for multiple reasons. First of all, the technology consistently yields excellent image quality. If you watch any of the DLSS-enabled demos in real-time, it's difficult to distinguish between native 4K with TAA and the same scene enhanced by DLSS.

Second, we're told that DLSS should only get better as time goes on. According to Nvidia, the model for DLSS is trained on a set of data that eventually reaches a point where the quality of its inferred results flattens out. So, in a sense, the DLSS model does mature. But the company’s supercomputing cluster is constantly training with new data on new games, so improvements may roll out as time goes on.

Finally, this is a technology that might be viable on entry-level Turing-based GPUs (as opposed to ray tracing, which requires a minimum level of performance to be useful), if those graphics processors end up with Tensor cores. We'd love to see low-end GPU play through AAA games at 1920 x 1080 based off of a 720p render.

[SteelCrysis]

https://www.tomshardware.com/news/rtx-20...37995.html

Product launches rarely go off without a hitch. But knowing that provides little comfort when you've spent $1,200 on Nvidia's flagship graphics card only to have it fail right away. That appears to be what's happening to numerous people who bought the RTX 2080 Ti at launch, forcing Nvidia to respond.

Digital Trends reported this week that several RTX 2080 Ti customers on Nvidia's support forum and Reddit have complained about various issues with their graphics cards. The exact problems vary: people have reported "crashes, black screens, blue screen of death issues, artifacts and cards that fail to work entirely," the report said. Nvidia's replacement cards seem to suffer from similar problems too.

https://www.extremetech.com/gaming/27976...high-rates

This is the type of issue that rapidly and understandably leads to angry users — nobody likes finding out that their high-end GPUs are non-functional from the factory, much less going through repetitive replacement cycles. So far, that’s the biggest red flag in all of this: Multiple reports from users claiming to be on their second RTX 2080 Ti, or to have had more than one card fail at the same time. That’s noteworthy because it implies that end users could be seeing much higher failure rates than normal. Anyone can get a single bad board, but receiving multiple bad cards in sequence can be a sign of larger problems.

https://www.techpowerup.com/249018/evga-...or-usd-999

Well this here is something that we don't see every day (read, never): an RTX 2080 Ti graphics card for $999. NVIDIA did announce pricing starting at that value for this particular graphics card, but pricing, as always with NVIDIA's Founder Editions, has always creeped towards the company's self-set $1,199. EVGA, however, has just put up a product page for their GeForce RTX 2080 Ti BLACK EDITION GAMING, a dual-fan solution (much like NVIDIA's own Founders Edition) with EVGA's iCX 2 cooling expertise that's being marketed at the unicorn-like $999 price-point, with a limit of 1 per household.

[SteelCrysis]

https://www.techpowerup.com/249273/gefor...n-the-rise

Multiple sources confirmed to GamersNexus that the GTX 1080 Ti is starting to be really difficult to find. Supplies are decreasing and the reason seems to be clear: NVIDIA could have stopped the production of those graphics cards. This has had an immediate effect on these cards' prices, which in the last few days have increased everywhere in the world. The performance differences with the new GeForce RTX 2080 are not that important if you don't need the RT part of the equation -we could confirm this on our own review-, but the price of these new graphics card have made considering a 1080 Ti a viable option for many users that are looking to upgrade their systems.
...
With prices climbing, some are claiming the same will happen to the GTX 1080, GTX 1070 or GTX 1070 Ti in the next few weeks. Reports of RTX 2080 and RTX 2080 Ti inexplicably dying on users could also be fueling consumer-fear, as well as a [temporary] erosion in the value proposition of the RTX 20-series itself, as Microsoft pulled Windows 10 1809 Update, leaving fewer people with DirectX Ray-tracing, the software foundation for RTX.

[SteelCrysis]

https://www.techpowerup.com/249345/nvidi...nd-2080-ti

Using today's 416.81 drivers I did a few quick runs for power consumption on all three existing GeForce 20 SKUs; the RTX 2070, RTX 2080 and RTX 2080 Ti.

As you can see, multi-monitor power consumption is finally back at normal levels. Good job, NVIDIA! While the power draw of Pascal cards is still a little bit lower, the differences are now negligible, and won't have any significant effect on power usage, power bill, temperatures, or the environment.

[SteelCrysis]

https://www.techpowerup.com/reviews/Perf...ing/5.html

DirectX Raytracing and NVIDIA RTX are as ambitious a technology as Ageia PhysX (which NVIDIA later acquired and popularized). Much like PhysX, DXR is facing a lot of teething problems with initial adoption. It requires you to not only have Windows 10, but also the latest "October Update". When we managed to get it off the ground, we were greeted by some astounding results. Performance of the flagship GeForce RTX 2080 Ti was cut by half in our first run at 1920 x 1080; we weren't aware DICE added another setting called "DXR reflections quality," which by default was set to "Ultra" (other settings include Low, Medium, and High).
...
For the RTX 2080 Ti, performance drops by over 58% just by flicking on DXR, and it does not improve at all, when switching from "Ultra" to "High" and "Medium." You only begin to see some performance improvements at the "Low" setting (44% performance impact). RTX impacts performance by over 50% at 1080p resolution, and the card only manages to keep frame-rates in the high 60s. Moving on to the 1440p resolution, and the trend of "Low" improving performance tangibly repeats, however, the overall frame-rates of "Ultra," "High," and "Medium" remain around 45 fps, and we're still only talking about the RTX 2080 Ti! Switch to the 4K UHD resolution, and you begin to realize the gravity of the situation. The game is barely playable, with twenty-something frames-per-second in DXR Ultra/High/Medium, and improves to 44 fps in the "Low" setting. Without RTX, we were cruising at 88 fps!

The GeForce RTX 2080 similarly loses 60% of its frame-rate at 1080p, as it drops from a 120 Hz-friendly 138 fps down to 54 fps, which barely keeps 60 Hz displays happy. At 1440p, frame-rates drop from 111 fps to around 39 fps, which is still playable, although a massive 64% performance loss. 4K UHD hovers around unplayable levels, such as 22 fps, which is downright unsuitable for online multiplayer gameplay. The RTX 2070 drops down to GTX 970 kind of frame-rates when RTX turned on. It is somewhat playable at 1080p, barely playable at 1440p, and practically unplayable at 4K UHD, which is highly disappointing as the RTX 2070 is being touted as the most popular SKU of its generation owing to its affordability.

Another thing that's problematic, specifically for Battlefield, is that DirectX Raytracing is built upon DirectX 12, a rendering mode that experiences some stuttering in Battlefield 5, something that was present even in BF1, and that the developer has been unable to completely fix.

[SteelCrysis]

https://www.extremetech.com/gaming/28065...ses-remedy

“Test escape” is a term of art that refers to a problem that slipped through quality control testing and wasn’t caught until it wound up in the hands of the end-customer. This certainly does appear to be the case. While I haven’t been running news up the flagpole every single time an RTX 2080 Ti failed, I promised to keep an eye on the situation and have been doing so. Kyle @ HardOCP has also done some of this work — one of the two RTX 2080 Ti GPUs he purchased failed after just two hours last Friday. Other cards have literally burst into flames and there’s been speculation in the GeForce forums that the problem might lie with Micron’s GDDR6. Supposedly cards coming back from Nvidia have Samsung memory instead of Micron.

A memory swap could also simply be evidence that Nvidia’s repair shop / GPU supplier had Samsung memory in-stock instead of Micron. Just because a company is using two different companies for supply doesn’t mean it always has an equal and identical stock of parts from both manufacturers on hand literally every single moment of the day, and we’d need to know more than we do about what issue Nvidia has identified and how it slipped through the cracks to know what happened here. Thus far, that information hasn’t been disclosed.
...
The optics of shipping defective hardware immediately after you jacked up the price on your flagship GPU by $500 are not good, and gamers who promptly sold their old cards upon purchasing new ones may find themselves stuck on Intel integrated graphics while the replacement goes through, but Nvidia has pledged to make the situation right for affected customers. What Nvidia needs to do now, if at all possible, is release information that will help gamers identify if they have a defective GPU before it fails, in order to request replacements as quickly as possible and avoid potential system damage.

[SteelCrysis]

https://www.tomshardware.com/news/nvidia...38083.html

Nvidia announced its third quarter financial results, but a worrying glut of graphics cards in the channel, which Nvidia CEO Jensen Huang chalked up to a "Crypto hangover," impacted overall earnings and guidance. The news sent the company's stock plummeting 18% in after-hours trading. Nvidia expects the oversupply to require "one to two" quarters to recede.

The oversupply specifically impacts Nvidia's GTX 1060 graphics cards, which could result in a delay in shipments of Nvidia's forthcoming Turing 2060 cards because Nvidia will surely wait for the existing stock to be liquidated before it floods the market with newer, faster video cards in the same price range.

AMD also released the RX 590 yesterday, which it designed specifically to compete with the GTX 1060. That could complicate matters for Nvidia as it tries to liquidate the excess inventory. The inventory announcement overshadowed the record revenue generated by several of the company's key segments, like data center and automotive.

[SteelCrysis]

https://www.extremetech.com/gaming/28080...tory-fears

While I don’t believe it played any part on the company’s share price slide, I can’t help but note that Jen-Hsun Huang fundamentally mischaracterized Turing in his remarks and Q&A. According to Jen-Hsun:

the Turing launch happened towards the end of the quarter, and it’s the biggest generational leap we’ve ever had… At every single price point it serves, it is substantially higher performance than the last.

This is an extremely stretched reading of the actual situation on the ground. Yes, the RTX 2070 is faster than the GTX 1080, the RTX 2080 is faster than the GTX 1080 Ti, and the RTX 2080 Ti is actually substantially faster, with a $500 price “upgrade” stacked on top. Since Nvidia decided to jack up the prices on these cards, none of them do a particularly good job justifying their pricing or price/performance ratios. That’s the entire problem. The marketing for these cards is writing checks that their actual performance can’t cash.

Investors, however, almost exclusively focused their questions on the channel inventory issue, which Nvidia said will take 1-2 quarters to completely clear. This has brought down their overall revenue estimates to $2.7B for Q4, which is significantly lower than what investors expected. Given that we heard rumors that Nvidia had to eat a substantial amount of inventory it’s not surprising to see this happen, and of course, there will be people who theorize that the problem is Turing, not Pascal. Nvidia’s official line is that Turing is selling very well and performing excellently, with no mention of the RTX 2080 Ti failures by themselves or analysts. 7nm ramps were not discussed, though Jen-Hsun specifically stated he wanted to have midrange cards in-market as quickly as possible.

[SteelCrysis]

https://www.tomshardware.com/reviews/bat...,5911.html

While the game was benchmarking on our test bench, we also played through some of the campaign on the Corsair Vengeance 5180 Gaming PC, which boasts an RTX 2080. We ran the game at the same FHD resolution with Ultra presets and medium Reflection Quality.

The first time you see ray tracing and know it, it’s pretty cool. You can see fire explode against the ground in the Tirailleur mission as planes drop bombs on your caravan. In Nordys, you can see the mountains reflected in the far side of a puddle in a pipe you're crawling through. And the reflections shift with your movements. Light shifts on sand as you drive a tank through Libya in My Country Calling. It’s really cool to see what the new tech can do.

But that’s when we were looking for it. That’s not to say we didn’t find it naturally sometimes, but when playing the game’s campaign or multiplayer while trying not to get shot, higher frame rates and resolution offer more benefit to the game than seeing light from an explosion bounce off the side of a tank.

Of course, we do know other games will use ray tracing in other ways. Shadow of the Tomb Raider, for instance, will use it to render more lifelike light and shadows. Developers will have to come up with a good combination of shadows, reflections and other ray-traced effects to make the performance hit worth it.

For competitive play, in Battlefield V right now? Play without ray tracing and get higher frame rates. It’s a better experience.

[SteelCrysis]

https://www.gamersnexus.net/guides/3394-...ck-screens

Just for good measure, we also took apart several cards and did a cursory board-level inspection. We were only really looking for anything extremely out of place, like missing thermal pads, poor contact, burned or damage components, and so on. There was only one device that demonstrated any physical defect, but it was unrelated to the issue of artifacting and is something we may discuss later. The thermal pad contact on all of the cards looks fine – we can see indentations and clear contact being made to the pads – and there are no obviously damage components. Any defect is going to be something we don’t have the tools or knowledge to see, as it’s likely in the board or in the silicon.

What we’ve done today is primarily rule things out – or mostly out, as we can’t speak with 100% certainty – like thermals, firmware, and Windows. We’ve also determined that blue screens were a separate issue, caused by early driver compatibility problems with many GSync monitors. Visible defects on the boards were not present, although we have no means of inspecting the internals of the board or the silicon.

Ubuntu Linux also exhibits artifacting issues (410 and 415), as does the newest firmware revision on the bad cards, and artifacting still occurs even when a water block is on the cards. We have to assume that thermals are not the issue on the cards we had. The biggest take away is that there’s not some magical “TjMax” trip-point whereupon artifacting kicks in. Even running our known-good review sample at 100+ degrees on the memory – decidedly a bad thing for the memory’s health – we still did not encounter a hard shutdown, thermal fail-safe, or artifacting. For anyone who thought artifacting was triggered upon hitting a certain threshold, it would appear that this isn’t the case. Sticking thermocouples on the cards also did not produce offensively high thermal numbers, although I1 was getting bad.

Speaking with some engineers in the industry, we’re left with an assumption that this is either a board-level assembly issue or an in-silicon issue. NVIDIA has noted that it is replacing any affected cards, so there’s that. The company described these units as “test escapes” and seems to think that the problem is fixed going forward. We have exceedingly high confidence that this is a hardware-level issue, but it does not appear to be an epidemic.

[SteelCrysis]

https://techreport.com/review/34267/exam...efield-v/6

Despite the enormous hit to average frame rates, it's commendable that RTX effects don't result in horribly uneven or jerky frame delivery from Battlefield V. So long as gamers with GeForce RTX cards can tolerate 1920x1080 or 2560x1440 gaming and less-fluid frame rates overall, our results suggest they'll still be able to enjoy consistently-paced frames that help maintain a sense of immersion.

On the whole, I don't think gamers who already have high-end graphics cards with performance comparable to an RTX card should ditch their current pixel-pusher and go buy one of Nvidia's latest on account of Battlefield V's reflections alone. We're in the early days of this technology, and the effects I witnessed probably have room for improvement as developers get their coding fingers around real-time ray tracing and the hybrid rendering approach that Nvdia's Turing cards require.

Still, if you were planning to upgrade to an RTX 2070, RTX 2080, or RTX 2080 Ti to begin with, you'll be happy to have RTX reflections in your load-out when you fire up Battlefield V, and after this first taste of the tech, I'm eager to see what other studios can do with Nvidia's hybrid-rendering tool kit for shadows, global illumination, ambient occlusion, and more. I'm hopeful it won't be a long wait to find out.

[SteelCrysis]

https://techreport.com/review/34269/test...colossus/3

Overall, content-adaptive shading is another intriguing Turing technology that seems to be in its infancy. All three of the Turing cards we have on the bench so far aren't lacking for shader power, and they're plenty capable of running Wolfenstein II at impressive frame rates even at 4K with maximum settings to start with. We're curious what CAS might do for potential lower-end Turing cards as a result of this testing, but for now, the tech is simply making great performance a little bit better. If you haven't played Wolfenstein II through yet, or at all, on a Turing card, you can leave CAS enabled without any worries and enjoy its minor performance-improving benefits.

[SteelCrysis]

https://www.tomshardware.com/news/nvidia...38116.html

As with all leaks, it's healthy to take the results with just a pinch of salt. The GeForce RTX 2060 used in the Final Fantasy XV benchmark was probably an engineering sample and subject to further tweaks and changes. So, don't jump to conclusions just yet. At the same time, we can't hide the excitement of seeing what it can do compared to the current generation of Nvidia and AMD graphics cards.

The GeForce RTX 2060 was reportedly benchmarked at the 3840 x 2160 resolution with the High Quality preset. According to the Final Fantasy XV benchmark scoreboard, the purported GeForce RTX 2060 put in a score of 2,589 points. It's up to 30.43 percent faster than the last-generation GeForce GTX 1060 6GB, which is a pretty big performance jump. The GeForce RTX 2060 even manages to close in on the GeForce GTX 1070 with the latter only being around 6.18 percent faster.

Things get heated when the GeForce RTX 2060 is placed against its AMD rivals. It reportedly performs up to 22 percent faster than AMD's latest Radeon RX 590 graphics card. Nevertheless, the GeForce RTX 2060 falls behind the AMD Radeon RX Vega by 4.44 percent. Unfortunately, the entry doesn't specify whether it's the RX Vega 56 or the RX Vega 64.

[SteelCrysis]

https://www.tomshardware.com/news/nvidia...38183.html

It doesn't come as a surprise that Nvidia is preparing the Turing iteration of its Titan graphics card, but it is shocking that the card is already out of the oven. Nvidia has already sampled it to popular YouTube personalities such as JayzTwoCents, Linus Tech Tips, Gavin Free from The Slow Mo Guys, and Andrew Ng, the founder and CEO of Landing AI.

[SteelCrysis]

Titan RTX announced, coming out later this month for $2,499: https://www.techpowerup.com/250191/nvidi...-usd-2-499

[SteelCrysis]

New patch for Battlefield V that promises up to 50% performance improvements with DXR: https://www.techpowerup.com/250201/dice-...-fps-gains

[SteelCrysis]

https://www.tomshardware.com/news/battle...38188.html

Updated, 12/4/2018, 7 a.m. PT: DICE announced that it's delaying the Battlefield V: Tides of War Chapter 1: Overture update for an undetermined time. The developer said it discovered some issues with the update and "rather than create issues in the game," it decided to hold it back until those problems are resolved. DICE said it doesn't foresee a long delay and will provide more information as soon as it becomes available, possibly today.

[SteelCrysis]

The Battlefield V update is up now: https://www.techpowerup.com/250259/battl...d-live-now

[SteelCrysis]

https://www.tomshardware.com/news/nvidia...38283.html

Nvidia may split its gaming GPU lineup into RTX and GTX series, including a new GTX 1160 Ti, according to reports. Videocardz suggests that a new card, the 1160 Ti, may launch around the same time as the rumored RTX 2060.
...
Additionally, the site says that there will be no RTX 2050, and that entry level cards will all be part of the 11 series.

[SteelCrysis]

https://www.techpowerup.com/250924/nvidi...e-and-type

There are at least two parameters that differentiate the six (that we know of anyway): memory size and memory type. There are three memory sizes, 3 GB, 4 GB, and 6 GB. Each of the three memory sizes come in two memory types, the latest GDDR6 and the older GDDR5. Based on the six RTX 2060 variants, GIGABYTE could launch up to thirty nine SKUs. When you add up similar SKU counts from NVIDIA's other AIC partners, there could be upward of 300 RTX 2060 graphics card models to choose from. It won't surprise us if in addition to memory size and type, GPU core-configurations also vary between the six RTX 2060 variants compounding consumer confusion. The 12 nm "TU106" silicon already has "A" and "non-A" ASIC classes, so there could be as many as twelve new device IDs in all! The GeForce RTX 2060 is expected to debut in January 2019.

[SteelCrysis]

https://www.extremetech.com/gaming/28283...ay-tracing

But there’s also another possible outcome here. I want to stress that this is entirely speculation on my part, but it’s possible that Nvidia intends to market RTX as an up-market feature in its own right, similar to how Nvidia has G-Sync displays and has previously sold features like PhysX. In this scenario, the GTX brand becomes the ‘lower-end’ branding on a longer-term basis with a variable crossover point depending on where the sweet spot is. In other words, we might have a GTX 1160 and RTX 2060 that offered equivalent rasterization performance, but with the 1160 occupying a lower price point and the RTX 2060 offering what Nvidia believes are forward-looking features.
...
There’s a big downside to this argument that I’ll go ahead and acknowledge: It would make it even harder to push RTX into the mainstream of GPU technology. Right now, the only known games planning to include RTX are Nvidia launch partners. Microsoft’s DXR standard for ray tracing is now part of DirectX, but if you consider the wider market including consoles — and game developers absolutely do when they decide which features to build into games — it’s mostly a market without ray tracing. None of the rumors around Navi suggest it supports ray tracing with specialized hardware, and the compute costs of the feature are high enough that it’s unlikely to be a major focus without additional dedicated resources.
...
I’d argue that the second approach — RTX as a premium feature for those willing to pay — actually aligns better with the way Nvidia has historically rolled out features. Even the argument that this exposes the company to weakness if a competitor comes along with a cheaper part is only as true as the presence of a competitor to launch such a GPU in the first place. With AMD currently fielding nothing to compete against Turing and Intel’s GPU efforts likely a year away, Nvidia may have seized on its current position at the top of the stack to introduce premium features as a short-term profit driver in a manner that it would never have attempted had AMD been offering better competition in the first place.

Which is true, if any? We don’t know yet. But with the RTX 2060 reportedly arriving in January, we’ll know more soon.

[SteelCrysis]

https://www.tomshardware.com/news/nvidia...38301.html

Thanks to Lenovo's recent leak, we can be fairly certain that the GeForce GTX 1160 will be offered in 3GB and 6GB memory configurations. This shouldn't really come as a surprise as Nvidia released the previous GeForce GTX 1060 with 3GB and 6GB as well, and then later launched a region-specific variant with 5GB.

[SteelCrysis]

https://www.techpowerup.com/251048/nvidi...red-tested

Here are some of the first pictures of NVIDIA's upcoming GeForce RTX 2060 Founders Edition graphics card. You'll know from our older report that there could be as many as six variants of the RTX 2060 based on memory size and type. The Founders Edition is based on the top-spec one with 6 GB of GDDR6 memory. The card looks similar in design to the RTX 2070 Founders Edition, which is probably because NVIDIA is reusing the reference-design PCB and cooling solution, minus two of the eight memory chips. The card continues to pull power from a single 8-pin PCIe power connector.
...
In its Reviewer's Guide document, NVIDIA tested the RTX 2060 Founders Edition on a machine powered by a Core i9-7900X processor and 16 GB of memory. The card was tested at 1920 x 1080 and 2560 x 1440, its target consumer segment. Performance numbers obtained at both resolutions point to the card performing within ±5% of the GTX 1070 Ti (and possibly the RX Vega 56 from the AMD camp). The guide also mentions an SEP pricing of the RTX 2060 6 GB at USD $349.99.

[SteelCrysis]

https://www.techpowerup.com/251108/gddr6...than-gddr5

The latest GDDR6 memory standard, currently implemented by NVIDIA in its GeForce RTX 20-series graphics cards, pulls great premium. According to a 3DCenter.org report citing list-prices sourced from electronics components wholeseller DigiKey, 14 Gbps GDDR6 memory chips from Micron Technology cost over 70 percent more than common 8 Gbps GDDR5 chips of the same density, from the same manufacturer. Besides obsolescence, oversupply could be impacting GDDR5 chip prices.

Although GDDR6 is available in marginally cheaper 13 Gbps and 12 Gbps trims, NVIDIA has only been sourcing 14 Gbps chips. Even the company's upcoming RTX 2060 performance-segment graphics card is rumored to implement 14 Gbps chips in variants that feature GDDR6. The sheer disparity in pricing between GDDR6 and GDDR5 could explain why NVIDIA is developing cheaper GDDR5 variants of the RTX 2060. Graphics card manufacturers can save around $22 per card by using six GDDR5 chips instead of GDDR6.

[SteelCrysis]

https://www.tomshardware.com/reviews/nvi...60-10.html

RTX 2060 lands in more hotly contested territory, though. AMD’s Radeon RX Vega 56 can conceivably compete with a lower price, while Radeon RX Vega 64 demonstrates similar performance. Plenty of GeForce GTX 1070 and 1070 Ti cards vie for attention too.

In short, it’s not enough for GeForce RTX 2060 to replace a Pascal-based card at the same price, add RT cores and tell enthusiasts that the games are coming soon. No, GeForce RTX 2060 needs to be faster and cheaper than the competition in order to turn heads.

A price tag of $350/£330 puts GeForce RTX 2060 in the same territory as GeForce GTX 1070. It’s less expensive than AMD’s Vega 56 and Nvidia’s 1070 Ti. Yet, it beats both cards more often than not. The geometric mean of RTX 2060’s average frame rate across our benchmark suite at 2560x1440 is 77.9 FPS. Apply the same calculation to GTX 1070 Ti and you get 76.2 FPS. RX Vega 64 achieves 77.8 FPS. RX Vega 56 sits at 69.8 FPS. GTX 1070 lands just under that, at 67.2 FPS.

Nvidia’s biggest sin is probably calling this card a GeForce RTX 2060. The GeForce GTX 1060 6GB launched at $250. GeForce GTX 960 started at $200. GeForce GTX 760 debuted at $250. Now, the company is pushing its xx60 series all the way up to $350. The performance we measured certainly justifies such a price. But it probably could have been called a 2060 Ti or the 2070 and made fewer waves.

The other interesting take-away from the launch is that Nvidia’s hybrid rasterization/ray tracing approach is still viable down at the 2060’s price point. As far back as our first deep-dive into the Turing architecture, we wondered how useful 36 RT cores would be on TU106 compared to TU102’s 68 RT cores. Now, we have a derivative GPU with just 30 RT cores, and it’s capable of over 60 FPS at 1920x1080 with all options, including DXR Reflection Quality, set to Ultra in Battlefield V. No doubt, that’s a testament to EA DICE and its optimization efforts, which continue in the form of an upcoming patch to enable DLSS support.

Still, we don’t draw conclusions based on what might happen down the road. Fortunately for Nvidia, RTX 2060 is generally faster than much more expensive cards in today’s games. Its 160W TDP does correspond to that higher performance. But it’s also still significantly more efficient than AMD’s Vega 56. We’re relatively confident that RTX 2060 Founders Edition, specifically, will see limited availability on geforce.com. Once it’s gone, Nvidia’s board partners need to keep prices close to the $350/£330 benchmark or else risk being undercut by very real competition from AMD and Nvidia’s previous generation.

[SteelCrysis]

https://www.extremetech.com/gaming/28326...060-at-349

Comparing the GTX 1060 to the RTX 2060, it’s hard not to feel like Nvidia is phoning in the specs a bit. While it’s true that the RAM loadouts on the RTX 2070 and 2080 stayed the same as their Pascal counterparts, the GTX 1060 was a thoroughly midrange card, with a 3GB version at ~$200 and a 6GB variant starting around $250. The RTX 2060 is a $349 card — and Nvidia was more than happy to sell you 8GB of VRAM in a $349 back in 2016 with a 256-bit memory bus attached. With AMD now selling 8GB GPUs for under $200, Nvidia’s decision to keep to a 6GB frame buffer doesn’t seem to necessarily bode well for the GPUs long-term future, especially given that next-generation consoles are expected to arrive in the next 12-20 months with 12-16GB of RAM.
...
The central problem with the RTX family from the beginning has been that it offers slim-to-no performance improvements over the GTX cards that were available at the same price. This GPU family is essentially a forward-looking feature play and as we’ve previously stated, we don’t ever recommend buying GPUs based on promises of support in future titles. If we assume that Turing will be Nvidia’s top-end GPU family for 22 months (Pascal set a record at 28 months), 3.5 of those months are already gone. That’s 16 percent of the GPUs entire life at the top of the product stack and there’s currently just one RTX-enabled game you can play. If you ever wanted a nutshell mathematical reason for why ExtremeTech doesn’t recommend buying GPUs today for features you won’t be able to use until tomorrow, there it is.

Compared with the RTX 2070, however, the RTX 2060 is a fairly good deal (Anandtech notes that it offers 86 percent of the performance at 70 percent the price). It improves on the GeForce 1060 6GB by 1.59x while costing 1.4x more. There is, at least, a better argument for the RTX 2060 in certain respects than there was for its higher-end cousins, but the reduced RAM loadout compared with previous generation Nvidia cards could prove a longer-term problem (Anandtech thinks they may have seen some evidence of a RAM-related bottleneck in Wolfenstein 2). Features like HDR and ray tracing itself can also increase RAM buffer pressure. We’re not saying the RTX 2060 will have a problem here, but it’s something to consider.

One difference between the RTX 2060 and the rest of the Turing stack is that in this case, AMD has GPUs competing in this space. The RTX 2060 isn’t just faster than Vega 56, it uses significantly less power while matching 95 percent of Vega 64’s performance. Up until today, AMD could argue that support for FreeSync made a difference in its overall value proposition, but Nvidia’s recent FreeSync support announcement will scythe that argument out from Team Red, assuming that NV support lives up to its promises. Either way, AMD now has a problem.

Right now, the RX 580 is $190, the RX 590 is $260, Vega 56 is $370, and Vega 64 is $400. If AMD wants to remain competitive on price/performance, the RX 590 is going to need to drop to something more like $220 – $240, Vega 56 needs to hit $300, and Vega 64 will need a $50 price cut as well.