Quote:That’s how we look at it. The RTX 2060 is better positioned, but much of the luster of a higher performance is stolen by higher price. It is a worthwhile buy in the use cases we’ve described above, which is far more than we could have said for the RTX 2080 at launch. RTX as a technology remains unproven at this point, with two loose implementations that leave much to be desired, so we’d advise against purchasing solely on the promise of wider-spread RTX and instead focusing on price-to-performance in traditional games. Game developers are a different story.

The RTX 2060 gets a soft recommendation from us, as its competition – even friendly fire competition – is multivarious and complicated. NVIDIA has done well to improve its market position with the RTX 2060. We are leaning positive on the RTX 2060, and moves to shift toward adaptive sync support (e.g. FreeSync) will yield good community response.

One more note that’s specific to the Founders Edition cooler: The cooler design is utterly, completely insane. NVIDIA seems lost in cooler design. The 2080 FE coolers used over 70 screws, if you remember (and we do), and the 2060 takes a similar approach. Curiously, the 2060 manages to do worse, namely by resorting to hot glue to hold the fan cable in place, not to mention the 7 screws used to further secure the fan cable. The PCIe connector is positioned right-side, past the PCB, which required an approach similar to the GTX 1060’s soldered power extension to the PCIe connector. Together, these two design approaches render the RTX 2060 Founders Edition card effectively impossible for a novice to service, e.g. for thermal paste replacement or dead fan replacement, and difficult for an expert to service. The card is becoming further Apple-like in design, and that’s not a good thing.

If buying the RTX 2060, we’d encourage board partners who know better how to build a cooler. We can, though, soft-recommend the RTX 2060, just take a good look at nearby price alternatives and think about your use cases. The RTX 2060 recovers NVIDIA's positioning for its RTX series and sets it back on track with a stronger product portfolio.

Quote:The RTX 2060 was rumored to come in half a dozen sub-variants based on memory size and type, although in the end, NVIDIA only launched the top-spec varient with 6 GB of GDDR6 memory. Perhaps, NVIDIA is saving the other SKUs up for when its GTX 1060 inventories are sufficiently off the shelves and spring-summer sets in.

Quote:Three industry sources, including one of Nvidia's board partners, told VideoCardz about the existence of the Turing-based GeForce GTX 1660 Ti graphics card. However, the final product name is still subject to change. In the meantime, the graphics card is being labeled as the GeForce GTX 1660 Ti.

The GeForce GTX 1660 Ti is reportedly the first Turing product to be released under the existing GTX branding. The GTX 1660 Ti is based on the TU116 die that TSMC produces with the 12nm FinFET manufacturing process. The card is said to lack ray tracing capabilities. In a perfect world, this should result in the graphics card having a lower price than its RTX brethren.

Quote:Nvidia CEO Jensen Huang recently told VentureBeat in a Q&A session that the Santa Clara chipmaker is clearing out the last of its GeForce GTX 10-series inventory. The high-end models, such as the GeForce GTX 1080 Ti, GTX 1080, GTX 1070 Ti, and GTX 1070 are already sold out with the GTX 1060 meeting the same fate in a few more weeks.
...
Currently, the GeForce RTX 20-series is a hard sell mostly due to its inflated price. Take the GeForce RTX 2060 for example. The performance is there as the graphics card performs faster than a GeForce GTX 1070 Ti. However, with a $350 price tag, the GeForce RTX 2060 doesn't feel like a mid-range product anymore. There are rumors that Nvidia could possibly launch the GeForce GTX 11-series which still use the Turing silicon but without the RT cores for ray tracing. If priced adequately, the GTX 11-series can be a viable option to the RTX 20-series.

Quote:These GPUs could still have tensor cores which are needed to accelerate DLSS, a feature highly relevant to this market segment.

Quote:The GeForce GTX 1660 Ti was tested in the popular Ashes of the Singularity (AotS) benchmark on the High quality preset at the 1920x1080 resolution. TUM_APÏSAK noted that this was the mobile version of the GeForce GTX 1660 Ti as the graphics card was housed inside a laptop alongside an octa-core Coffee Lake H (CFL-H) processor.

The mobile GeForce GTX 1660 Ti put up a score of 7,400 points, which makes it approximately 19.35 percent faster than the GeForce GTX 1060. The mobile Nvidia GeForce GTX 1060 scored 6,200 points in the same benchmark with the same settings.
...
According to what VideoCardZ has heard, Nvidia could announce the GeForce GTX 1660 Ti as soon as next month. There is a rumor that a GeForce GTX 1660, which uses GDDR5 or GDDR5X memory, is also in the works, and it might be unveiled at the same time frame.

Quote:

• We have independently confirmed parts of the rumors with sources in the industry
  
• GN confirmed that the product will indeed drop RT and Tensor cores while upgrading the rest of the CUDA cores to Turing
  
• GN confirmed that this isn’t just a straight Pascal refresh and rebrand, as we speculated about 8 months ago. It is on Turing silicon
  
• GN confirmed independently that the 1660 will be less powerful than the 2060 in ways beyond RTX performance, namely in reductions to SM count
  
• This is all interesting, as it begins to loosely put a price on RTX itself
  

Quote:Now, HardOCP's sources tell us that the Nvidia GeForce GTX 1660 Ti will launch on February 15, with an MSRP of $279. The GTX 1660 will launch in early March at $229, while the GTX 1650 will go on sale for $179 in late March. Additionally, Nvidia will continue to supply the 1050 TI to retailers, and its price will drop to keep it competitive. The exact launch prices are still in flux, and any performance figures that show up on the web need to be taken with a grain of salt until independent testers like us get the cards in our labs.

Quote:Graphics card manufacturer Palit recently registered various new products with the Eurasian Economic Commission (ECC), including graphics cards using Nvidia's rumored GeForce GTX 1160, GTX 1660 and GTX 1660 Ti. This new listing disputes rumors that the GeForce GTX 1160 is fake; although, Lenovo has previously listed the graphics card with its Legion laptops.

Quote:NVIDIA is in the process of rolling out the first implementations of its RTX 2000 series GPUs in mobile form, and if the going is as is being reported, it's going to be a little rough for users to actually extrapolate their performance from product to product. This is because manufacturers are apparently getting a whole lot of leeway in how to clock their products, according to their solution's thermal characteristics and design philosophy.

What this means is that NVIDIA's RTX 2080 Max-Q, for example, can be clocked as low as 735 MHz, which is a more than 50% downclock from its desktop counterpart (1,515 MHz). The non-Max-Q implementation of NVIDIA's RTX 2080, for now, seems to be clocked at around 1,380 MHz, which is still a close to 200 Mhz downclock. Of course, these lowered clocks are absolutely normal - and necessary - for these products, particularly on a huge chip such as the one powering the RTX 2080. The problem arises when manufacturers don't disclose clockspeeds of the GPU in their particular implementation - a user might buy, say, an MSI laptop and an ASUS one with the exact same apparent configuration, but GPUs operating at very different clockspeeds, with very different levels of performance. Users should do their due research when it comes to the point of choosing what mobile solution sporting one of these NVIDIA GPUs they should choose.

Quote:But back to point, as it stands right now the GeForce RTX 2060 has enough VRAM to power through today’s games using maximum quality settings. As a GPU targeting 1440p gaming or extreme high refresh rate 1080p gaming, it fits the bill nicely, at least in terms of performance.

We can all agree 8GB of GDDR6 memory would have been better down the road, but for those buying a new graphics card right now it seems like a non-issue. Rest assured, we'll be watching the RTX 2060 over the next few years and monitor it closely against its peers as we go.

Quote:In short, the data we’re going to examine should be treated as interesting and useful, but not the absolute final word on the topic.
...
I don’t want to read too much into these graphs. We have a limited number of Turing data points to pull from and only a few months of data to work with. But the pattern suggests that Turing sales are indeed lagging Pascal in ways that may not be entirely explained by either higher prices, previous models still being in-market, or a larger user-base of Steam gamers.

Turing adoption should improve now that Pascal cards are exiting the market, but the GPU family appears to be off to a slow start. All of this is in-line with Nvidia’s market warning, but it’s interesting to see what the differences might be. There is a little good news in the SHS for NV, however — the RTX 2070 posted strong growth, nearly doubling its market share. Other strong movers (in terms of share gain) include the GTX 1070 (up 0.18 percent) and AMD’s RX 580 (up 0.15 percent)

Quote:A Russian retailer has leaked more specifications of NVIDIA's upcoming GeForce GTX 1660 Ti graphics card. Based on the 12 nm "TU116" silicon, this card will be configured with 1,536 "Turing" CUDA cores, but have no RT cores, and hence no RTX features. The chip could end up with 96 TMUs and 48 ROPs. The GPU is clocked at 1500 MHz nominal, and the boost frequency is set at 1770 MHz, however, the latter could be a factory-overclock set by AIC partner Palit for their GTX 1660 Ti StormX graphics card.

The memory subsystem of the GTX 1660 Ti is interesting. While it's still 6 GB of GDDR6 memory across a 192-bit wide memory bus, the memory clock itself is lower than that of the RTX 2060. The memory ticks at 12 Gbps, which results in 288 GB/s of memory bandwidth, compared to the RTX 2060, which thanks to its 14 Gbps memory achieves 336 GB/s. The card draws power from a single 8-pin PCIe power connectors. Outputs include HDMI, DisplayPort, and DVI, we don't expect any cards to ship with VirtualLink.

Quote:In all, there are 9 SKUs for the GTX 1660 Ti 3 GB graphics card being filed with the EEC, which usually preempts graphics card launches in those domains. These slot in nicely with ASUS' plans for 6 GB versions of the GTX 1660 Ti, almost to a card - though ASUS' STRIXX-branded graphics cards seem, for now, to only be available in 6 GB versions. Of course, the 3 GB of VRAM on the GTX 1060 allow the card to achieve a desirable performance/dollar ratio, but at the cost of some performance, with the penalty increasing alongside resolution - but these are cards that likely won't ever be used for 4K gaming. While 3 GB graphics cards still fare relatively well, as we've seen, the latest games are pushing over 3 GB of video RAM more often than not, which leaves the 3 GB version of the graphics card somewhat of a less than choice when it comes to AAA gaming. But when it comes to competitive multiplayer game,s it likely will be more than enough.

Quote:Take Metro Exodus first, with the relevant notes in the first image below. DLSS can only be turned on for a specific combination of RTX GPUs ranging from the RTX 2060 to the RTX 2080 Ti, but NVIDIA appear to be limiting users to a class-based system. Users with the RTX 2060, for example, can't even use DLSS at 4K and, more egregiously, owners of the RTX 2080 and 2080 Ti can not enjoy RTX and DLSS simultaneously at the most popular in-game resolution of 1920x1080, which would be useful to reach high FPS rates on 144 Hz monitors. Battlefield V has a similar, and yet even more divided system wherein the gaming flagship RTX 2080 Ti can not be used with RTX and DLSS at even 1440p, as seen in the second image below. This brought us back to Final Fantasy XV's own DLSS implementation last year, which was all or nothing at 4K resolution only. What could have prompted NVIDIA to carry this out? We speculate further past the break.
...
Regardless of what is the underlying cause, all in-game DLSS implementations so far have come with some small print attached, that sours the ultimately-free bonus of DLSS which appears to work well - when it can- providing at least an additional dial for users to play with, to fine-tune their desired balance of visual experience to game FPS.

Quote:VideoCardz scored not just pictures of the ASIC, but also the PCB of an MSI GTX 1660 Ti Ventus graphics card, which reveals something very interesting. The PCB has traces for eight memory chips, across a 256-bit wide memory bus, although only six of them are populated with memory chips, making up 6 GB. The GPU's package substrate, too, is of the same size. It's likely that NVIDIA is using a common substrate, with an identical pin-map between the TU106 and TU116, so AIC partners could reduce PCB development costs.

Quote:Of course, AMD may only be speaking from the point of view of a competitor that has no competing solution. however, company representatives said that they could, in theory develop something along the lines of DLSS via a GPGPU framework - a task for which AMD's architectures are usually extremely well-suited. But AMD seems to take the eyes of its DLSS-defusing moves, however, as AMD's Nish Neelalojanan, a Gaming division exec, talks about potential DLSS-like implementations across "Some of the other broader available frameworks, like WindowsML and DirectML", and that these are "something we [AMD] are actively looking at optimizing… At some of the previous shows we've shown some of the upscaling, some of the filters available with WindowsML, running really well with some of our Radeon cards." So whether it's an actual image-quality philosophy, or just a competing technology's TTM (time to market) one, only AMD knows.

Quote:Nvidia presented its fourth quarter and fiscal 2019 earnings today, which reflected a surprising 45% year-over-year loss in gaming GPU revenue. Overall the company performed better than expected after it lowered its earnings projections last month, with the $2.21 billion fourth quarter beating expectations and leading to a jump in its share price in after-hours trading (despite the 24% year-over-year loss in quarterly revenue). The company also chalked up a solid 21% gain in full fiscal year revenue.

Nvidia blamed the severe gaming downturn on several factors, including that sales of its Turing RTX 2070 and 2080 graphics cards were below its expectations for the launch of a new architecture. Nvidia opined that gamers might be waiting for more real-world examples of ray-traced games, which were notably absent at the launch of the Turing graphics cards, before purchasing a new graphics card.

Nvidia CEO Jensen Huang also pointed out that, for the first time in the company's history, its newest architecture launched with only high-end models available. The early RTX 2070 and 2080 cards came as specialty high-end and overclocked models that retail well above MSRP, while the lower-priced models didn't make their way to market until months later, leading to stunted sales.

Nvidia also delayed the release of the mainstream RTX 2060 cards due to the oversupply of GTX 1060 cards in the channel, which was largely due to the collapse of the cryptocurrency market. "The inability to launch [the] 2060 was a big inhibitor for us, but we did so at CES," said Huang, citing that these models address the broadest swath of the gaming GPU market.
...
Nvidia's gaming segment comprises 43% of its total revenue, so the 45% year-over-year decline is troubling. The company's $1.98 billion in GPU revenue fell 20% on the year, primarily from the declining sales of gaming GPUs, but the company projects revenue to be flat or only slightly down for the full year.

Quote:NVIDIA is reportedly pricing the GeForce GTX 1660 Ti at USD $279 (baseline pricing), which implies pricing of custom-designed and factory-overclocked cards scraping the $300-mark. The card is also spaced $70 apart from the RTX 2060, which offers not just 25% more CUDA cores, but also NVIDIA RTX and DLSS technologies. In media reporting of the card so far, it is being compared extensively to the GTX 1060 6 GB, which continues to go for under $230. Perhaps NVIDIA is planning a slower non-Ti version to replace the GTX 1060 6 GB under the $250-mark. That entry would place three SKUs within $50-70 of each other, a tight squeeze.

Quote:The GTX 1660 Ti’s MSRP could be quite low considering that the iBuyPower listing has the entire PC, which also includes a Core i7-8700, 480 GB SSD, and 16 GB of RAM, priced at $1200. Similar PCs from iBuyPower featuring the RTX 2060 retail for roughly $100 more. HardOCP and later VideoCardz reported that the GTX 1660 Ti's MSRP is $279 ($70 lower than the RTX 2060).

Quote:As per notorious leaker TUM_APISAK's tip, Nvidia's upcoming GeForce GTX 1660 Ti has been benchmarked by an unknown source with Square Enix's Final Fantasy XV benchmark tool. The GeForce GTX 1660 Ti certainly appears to give the GeForce GTX Titan X a run for its money.

Quote:So this not only speaks to NVIDIA recognizing that DLSS image quality isn't at the level it's supposed to be (which implies it can actually degrade image quality, giving further credence to AMD's remarks on the matter), but also confirms that they're constantly working on improving DLSS' performance and image quality - and more interestingly, that this is something they can always change, server-side. I'd question the sustainability of DLSS' usage, though; the number of DLSS-enabled games is low enough as it is - and yet NVIDIA seems to be having difficuly in keeping up even when it comes to AAA releases. imagine of DLSS picked up like NVIDIA would like it (or would they?) and expand to most of the launched games. Looking at what we know, I don't even think that scenario of support would be possible - NVIDIA's neural network would be bottlenecked with all the processing time required for these games, their different rendering resolutions and RTX settings.

DLSS really is a very interesting technology that empowers the RTX graphicvs card of every user with the power of the cloud, as NVIDIA said it did. However, there are certianly some quirks that require more processing time than they've been given, and there are limits to how much processing power NVIDIA can and will dedicate to each title. That the network needs to be trained again and again and again for every new title out there bodes well for a controlle,d NVIDIA-fed games development environment, but that's not the real world - especially not with an AMD-led console market. I'd question DLSS' longevity and success on these factors alone, whilst praising its technology and forward-thinking design immensely.

Quote:The 1060 had a good run. It launched at $250 and served up excellent frame rates at 1920 x 1080, gingerly stepping on Radeon RX 480’s toes in the process. However, GeForce GTX 1660 Ti blows right past it in the benchmarks. Our results show the 1660 Ti averaging about 100 FPS across our suite, beating Radeon RX 590, roughly tying the old GeForce GTX 1070, and losing slightly to Radeon RX Vega 56. And that’s at a price point just $30 higher than the 1060 6GB in 2016.

Step up to 2560 x 1440 and the GeForce GTX 1660 Ti continues delivering playable performance with an average of more than 60 FPS between 12 different games. That’s 19% faster than Radeon RX 590 and darn near even with GeForce GTX 1070. Radeon RX Vega 56’s lead grows to 10%, likely due to a massive memory bandwidth advantage that staves off bottlenecks at higher resolutions. But the cheapest model in stock on Newegg at the time of writing was $400. A 43% price premium on a card rated for 75%-higher power consumption just doesn’t make sense for 10%-higher frame rates.

AMD does appear to be chipping away at Radeon RX 590’s price. Entry-level models still cost $260, though. They need to be lower to justify 85% of GeForce GTX 1660 Ti’s performance at 188% of its power consumption.

Previous-gen GeForce cards get their fair share of shade, too. Remaining GeForce GTX 1060s fail to shine with GeForce GTX 1660 Ti beating them by 30%+, despite the fact you can find a 6GB model in the $260 range. And obviously you’re going to want to avoid the few GeForce GTX 1070s left on shelves priced at $310 and up.

Taking a step back, then, it looks like GeForce GTX 1660 Ti is the card to beat for fast-paced gaming at 1920 x 1080 and solid performance at 2560 x 1440. Our only hesitation in recommending it comes from GeForce RTX 2060, which doesn’t look as good in our performance per dollar charts but does include Nvidia’s Tensor/RT cores.

Quote:GeForce GTX 1660 Ti is NVIDIA's answer for the highly competitive sub-$300 segment. The card is based on the all-new TU116 graphics processor, which has been specifically designed to meet the demands of that market, which is mostly "price". Unlike other Turing GPUs, TU116 does not feature acceleration for RTX real-time raytracing or DLSS, because the specialized hardware consumes a significant portion of the die area on other Turing GPUs, which increases manufacturing cost significantly. NVIDIA did keep the other improvements of Turing though; like GDDR6 memory, larger caches, concurrent execution of float and integer operations and adaptive/variable rate shading.

As a result, when averaged over all our gaming benchmarks at 1080p, we see GTX 1660 Ti beat the Pascal based GTX 1070, and roughly match AMD's RX Vega 56 — pretty impressive for a mid-range card. While we don't have a GTX 1660 Ti reference-design, we expect this card to perform very closely to one, because it is clocked only marginally higher than reference. Compared to the RTX 2060, which is NVIDIA's next-fastest SKU, the MSI GTX 1660 Ti Ventus XS is 13% behind. Compared to GTX 1060 6 GB, which the GTX 1660 Ti replaces, the performance uplift is 40%, at a higher price point though. AMD just released the RX 590, on a 12 nanometer process no less, to address the growing requirements of the mid-range segment, GTX 1660 Ti makes short shrift of that, offering 30% more performance. With those performance results, GTX 1660 Ti is a great choice for gamers with a Full HD monitor, running at maximum details. If you are willing to dial down detail settings a bit, then it should be able to reach 60 FPS at 1440p in most titles, too.
...
The lack of ray-tracing and DLSS on GTX 1660 Ti seems like a big deal at first, especially considering how much NVIDIA is pushing those technologies. While both are extremely promising, they are not the most important thing to have right now, especially when every dollar matters. While I have no doubt that RTX support will be growing vastly, only few titles support it at this time, so I don't think anyone could be blamed for wanting to skip the tech for now and wait until it matures more. Looking at silicon economics, including RTX/DLSS would have either driven die size (= cost) so high that reaching the targeted sub-$300 price point wouldn't be possible anymore, or the number of shader cores would have had to be reduced, which would have resulted in no performance improvement over AMD — NVIDIA went the other route. "RTX" is a great vehicle for NVIDIA to sell a feature, which is more than just "higher FPS". Specifically, this targets people who are comfortable with 1080p 60 FPS and are even willing to play low details, as long as their GPU is affordable. If however they feel they miss out on visuals, something that DX12 couldn't achieve, they might be willing to spend more and go for RTX 2060, for example. The next step for NVIDIA is transitioning to the 7 nm production process, which increases density and should allow smaller GPUs to have RTX, too. I also wouldn't be surprised to see NVIDIA haters on the forums, who previously dismissed RTX as "useless", now ask "wut no RTX?"

Priced at $279, the MSI GTX 1660 Ti Ventus XS comes at NVIDIA MSRP pricing, which is great, as it makes the card much more competitive against other offerings, be them from AMD or NVIDIA (Pascal). GTX 1660 Ti does compete with RX 590, which it beats with ease: 30% faster, a fraction of the power consumption, similar fan noise. The RTX 2060 starts at $350, is 15% faster, and offers NVIDIA's RTX features like ray-tracing and DLSS — definitely an option if you can save up more money and are looking to play at 1440p. Strong competition has come in the past few days from AMD Vega 56, which has seen its price dropped to $330 (€ 269 including VAT in some European countries). This of course makes a compelling argument for people looking for price/performance, but Vega 56 runs much hotter and noisier than GTX 1660 Ti. Looking at the board designs I have no doubt that prices for GTX 1660 Ti can go down much further, so I wouldn't be surprised to see it at or below $250 soon, at which point it would spark the upgrade-itch for many users, especially those with older cards, like GeForce 900 series.

Quote:Anandtech’s table sums things up nicely. The problem with Turing, up until now, is that Nvidia raised prices on every Turing card relative to its Pascal counterpart at the same position in the product stack. This forced gamers into a binary choice: You can buy an RTX card for roughly what you paid for your previous GTX card and enjoy no extra performance or you can spend more money for a performance jump.
...
Regardless of what AMD does, however, we’ve finally gotten a Turing GPU that offers a genuine improvement to both performance and performance-per-dollar. Apparently, the only thing that required was dumping the features that were supposed to make Turing shine in the first place. Depending on how you feel about those features, this may not even count as a negative.

Quote:According to DigiTimes’ sources, Nvidia will launch the GTX 1660 on March 15 and GTX 1650 on April 30, for a minimum price point of $229 and $179, respectively. However, the release date for the 1650 conflicts with a report from VideoCardz last week and HardOCP in January claiming the GTX 1650 will come out in March.
...
According to DigiTimes' sources, Nvidia will refocus its GPU business on the mid-end and low-end in the first half of this year. However, Nvidia seems to want to lower prices not just to steal market share from AMD, but also because it’s looking to clear its high inventory, caused by a shrinkage in demand in the cryptocurrency mining market. Furthermore, the lower price points may help stabilize the company's revenues, which have been in a declining trend lately.
...
According to DigiTimes’ sources, it will take another quarter for Nvidia to clear its old inventory, and the company’s revenues are expected to stay flat compared to the previous quarter. Market watchers estimated that Nvidia’s revenue in 2019 will fall by 15-20 percent.

Quote:Our first data point comes from the bestselling lists at Amazon and Newegg. At Christmas and before, these lists were dominated by GTX 1080 and 1070 cards. The steady release of new cards and the retirement of old ones have changed this, and now only a relative handful of lower-end Pascal cards hold the space. The RTX and new 1660 Ti are taking over the market, as predicted.

Steam’s Hardware Survey data isn’t perfect. 10-12 percent of the GPUsSEEAMAZON_ET_135 See Amazon ET commerce it surveys are listed as “Other” in the list, which means there’s a substantial amount of baked-in inaccuracy. It’s a trailing list, which means we see the data for the month previously, and we don’t know when GPUs accrue enough market share to claim space. The cut-off looks to be ~0.15 percent, based on our data.
...
Last month, we noted that adoption rates between the RTX 2080 and RTX 2070 were nearly identical. That’s no longer the case — the RTX 2070 has pulled significantly ahead of the RTX 2080, at 0.5 percent uptake. This compares poorly against the GTX 1070, but the price-equal comparison between the GTX 1080 and the RTX 2070 is a fairer way to compare the two, in our opinion. The GTX 1080 was still decisively ahead of the RTX 2070 at the same point in its life cycle, but the gap is much smaller.

We’ve added the GTX 1080 Ti and RTX 2080 Ti as well, and a comparison between the RTX 2080 and the GTX 1080 Ti. In this case, we’ve had to omit a month from the GTX 1080 Ti — the third month of adoption data for that GPU slams into the August 2017 discontinuity in the Steam Hardware Survey data. The first two months of data show much stronger adoption rates for the 1080 Ti compared to either of the two Turing cards, though the RTX 2080 is in a stronger position.

It is not clear how to split the issues with Turing’s slow adoption against the hammering the GPU market took in the back half of 2018 in general. You can read the slow Turing sales as significantly caused by the slowdown, or argue that the slowdown has been as bad as it is in part because NV raised Turing prices, making Pascal more attractive (possibly as a means of clearing inventory).

Quote:Both the 1660 and 1650 Ti are expected to use GDDR5, not the faster, more expensive GDDR6 that debuted with upper-end GPUs in the product stack. The GeForce 1660 should be competition for the RX 590, given that it increases core counts by ~10 percent over the GTX 1060. Overall power consumption will undoubtedly favor Nvidia in this comparison, whether the RX 590 or RX 580 is the reference target. We’ll reserve thoughts on absolute performance until we see shipping data, though we’d expect broad performance parity. Power on the 1660 looks to be provided by a single 8-pin PCIe port.
...
Assuming these are the configurations Nvidia brings to market, it doesn’t look as though the company intends to particularly challenge AMD below the $200 price point. To be fair, there isn’t much sign it needs to do so. The latest figures from Jon Peddie Research show AMD losing significant market share in Q4 2018 compared with Q3, and performing even more poorly against Q4 2017.

Quote:If you’re gaming at 1920 x 1080, the 1660 Ti’s speed may seem a bit superfluous in light of the vanilla 1660’s smooth frame rates at an even more attractive $220/£200 price point. In fact, according to our calculations, GeForce GTX 1660 would seem to take the crown in a comparison of average frame rate per dollar, usurping GeForce GTX 1660 Ti, the respectable GeForce GTX 1070 Ti, and GeForce RTX 2060.

But does it really? When we fold in benchmark results from the older Radeon RX 580, generated by Tom’s Hardware Germany using a similar test system, and then apply U.S. pricing, AMD’s card scores just a bit better. Granted, it gets wrecked when we switch to a performance per watt comparison. If you’re not really concerned about power consumption, though, it’s really hard to pass on an 8GB Radeon RX 580, currently available at $190, for 1080p gaming.

Even if AMD’s Radeon RX 580 blowout comes to an end, the GeForce GTX 1660 will remain a solid pairing with high-refresh FHD displays. It’s less QHD-friendly than GeForce GTX 1660 Ti, and we’d suggest stepping up to at least a GeForce RTX 2060 if you’re playing at 2560 x 1440. However, when push comes to shove, you can dial down your favorite game’s detail settings to achieve respectable frame rates at higher resolutions on GTX 1660, too.

We approved of GeForce GTX 1660 Ti for its GeForce GTX 1070-like performance at a price point under $300 (and 120W power consumption). Similarly, we like that GeForce GTX 1660 drives great 1080p frame rates at $220. And we’re even more excited that there’s a competing card from AMD with real value chops we can mention in the same breath.

Quote:With GeForce GTX 1660, NVIDIA is clearly attacking AMD's Radeon RX 590 offering. The GTX 1660 is based on the same TU116 graphics processor that we saw on the GTX 1660 Ti not long ago. Besides shader count and clock frequencies the biggest difference between both cards is certainly that GTX 1660 uses GDDR5 memory whereas the GTX 1660 Ti uses GDDR6. Unlike other Turing GPUs, TU116 does not feature acceleration for RTX real-time raytracing or DLSS because the specialized hardware consumes a significant portion of the die area on other Turing GPUs, which increases manufacturing cost significantly. NVIDIA did keep the other improvements of Turing though, like larger caches, concurrent execution of float and integer operations, and adaptive/variable rate shading.

As a result, when averaged over all our gaming benchmarks at 1080p, we see the EVGA GTX 1660 XC Ultra beat AMD's Radeon RX 590, by a solid 11% margin. The GeForce GTX 1660 Ti and last generation's GTX 1070 are 12% faster. With 20% higher performance than GeForce GTX 1060 6 GB, at slightly higher pricing, the GTX 1660 is in a position to conclusively replace the GTX 1060 easily. EVGA gave their card a mild overclock, to 1845 MHz rated Boost, which is a 60 MHz increase — not a lot. With those performance results, the GTX 1660 is a good choice for gamers running high details with a 1080p Full HD monitor. Compared to GTX 1660 Ti you do have to sacrifice a little bit of settings to reach 60 FPS, but you'll save good money while doing so.
...
Compared to Pascal, Turing improved power-efficiency once more, and the GTX 1660 is no exception. EVGA's GTX 1660 XC Ultra uses only 110 W during gaming, which is even more impressive when you consider how much gaming performance the card has in it. Even compared to the latest Turing RTX GPUs, the GTX 1660 achieves 10% better performance per watt. Against AMD's Radeon RX 590, which was just recently released and uses a 12 nm production process too, the GTX 1660 is 2.5x more power-efficient; that's 250%!
...
It looks like the aging GTX 1060 6 GB has finally found a worthy successor. The GTX 1660 offers better performance and efficiency, at a similar price point — NVIDIA will sell millions of these. AMD's Radeon RX 590 will be a tough sale now. GTX 1660 is faster, cooler, quieter and cheaper — all at the same time. The only win for RX 590 is that it comes with a strong three-game bundle, which could offset your cost significantly. NVIDIA doesn't bundle any games with GTX 1660, but EVGA has stepped up and is offering the Grip racing game with their GTX 1660. I have no doubt that AMD will quickly adjust their pricing to address the changed market conditions. Radeon RX 570 and RX 580 have already seen price drops, down to $130 and $170, respectively. Especially RX 570 is now at a price point that is so low that many people are going to be willing to overlook its shortcomings both in performance and noise levels — and the two-out-of-three game bundle will help here there, too.

Quote:On a call with reporters, Nvidia was coy about what exactly the GTX cards will be capable of. It was suggested that a game like Battlefield V, which uses ray tracing for reflections on only some surfaces, would run far better than Metro Exodus, which utilizes a heavier workload to generate global illumination effects. It did suggest that lower-level DXR presets would play better with GTX, while RTX would be necessary for higher presets.
...
Nvidia added that opening up ray tracing to these cards was always the plan and that the company’s main goal has been to spread ray tracing to as wide an audience as possible. It’s unclear if this is at all a response to the RTX line up’s high pricing and the response in the marketplace.
...
Not all of the RTX series’ special sauce is spilling over to the GTX line with the April driver. DLSS, for example, is still locked to those cards because they require Nvidia’s Tensor cores.
...
Nvidia also announced that DXR is coming to the Unreal Engine and Unity, two of the biggest engines out there, allowing many studios to get easier access to ray tracing effects.

Quote:In fact, Nvidia’s entire blog post seems mostly calculated to persuade people not to run ray tracing workloads on GTX GPUs. The company takes multiple opportunities to note that the ray tracing portion of the rendering algorithm runs up to 10x faster on RTX cards, and that by enabling DLSS, RTX 2080 GPUs are up to 3x faster than GTX 1080 Ti GPUs.

According to Nvidia, attempting to run ray tracing on a GTX GPU will send the frame rate cratering, all the way down to 18fps on a GTX 1080 Ti. Of course, there are a few caveats to that. First, the 1080 Ti is using 1440p, despite the fact that 1080p was a target resolution for the RTX 2080’s ray tracing support to start with. Second, the 1080 Ti is using ray tracing set to Ultra. These aren’t the comparison points you’d use if the goal was to showcase that the 1080 Ti can deliver ray tracing. They’re the comparisons you use to show how the GTX 1080 Ti can’t actually deliver ray tracing, because you want to push gamers towards buying RTX cards.
...
Nvidia’s efforts to paint RTX as an unvarnished breakthrough for GPU rendering have foundered on the rocky shoals of objective reality. But game developers won’t adopt a feature if there’s no support for it, and support is built through GPU sales — sales which, according to the Steam Hardware Survey, are sharply off Pascal adoption rates last generation. So Nvidia is hoping that by seeing RTX support on GTX cards, consumers will respond more strongly.
...
Nor does Nvidia have any reason whatsoever to provide them. The point is to push players into upgrading, not offer acceptable performance on older hardware. This is not to say that Nvidia would take steps to cripple the performance of ray tracing on older GPUs, but they certainly don’t have any reason whatsoever to optimize it.

Quote:A Eurasian Economic Commission filing revealed many more details of this card, as an MSI Gaming X custom-design board was finding its way through the regulator. The filing confirms that the GTX 1650 will pack 4 GB of memory. The GPU will be based on the new 12 nm "TU117" silicon, which will be NVIDIA's smallest based on the "Turing" architecture. This card will likely target e-Sports gamers, giving them the ability to max out their online battle royale titles at 1080p. It will probably compete with AMD's Radeon RX 570.

Quote:A new patch has become available for Shadow of the Tomb Raider, which updated the game to the latest graphical technologies in the form of RTX and DLSS. The PC port of the game has been handed by developer Nixxes, which partnered with NVIDIA to work on adding ray-tracing enabled shadows to the game (there's a thematic coherence there if I've ever seen one).

Quote:Performance cost of RTX raytracing is comparable to other titles and in the 30-50% range depending on the actual scene, which is definitely not negligible. Also, the game has a few bugs related to settings, and the performance difference between "RTX High" and "RTX Ultra" is almost non-existent. This is similar to what we've seen in Battlefield V's first version of RTX raytracing and suggests that Eidos still has some optimizing to do. In Battlefield, developer DICE was able to gradually reduce the RTX performance hit over several patches (NVIDIA's developer relations helped a lot here). Given how important RTX is to NVIDIA, I have no doubts we'll see the same happen in Shadow of the Tomb Raider. As mentioned before, I'd consider this more of a test run for future titles and to introduce developers to the tech, so they can create great things in their next game—the learned knowledge won't be lost.

DLSS, on the other hand, does very well in Shadow of the Tomb Raider. We see around 35% performance gained at 4K, which is quite significant and nearly offsets the performance hit from RTX. Image quality seems very decent. I am not sure if this due to good DLSS training data or because the art style of Shadow of the Tomb Raider tends to work better with DLSS. Of course, if you look at individual pixels and do side-by-side image comparisons, you'll always be able to make out areas where DLSS looks blurry. Personally, I find DLSS-on vs. DLSS-off really hard to tell apart, which made organizing my screenshots a bit difficult because the differences are small. Just like in previous versions of the tech, you are free to turn on either DLSS or RTX, or both, as suits your taste and performance requirements.

Quote:The Final Fantasy XV results showed that the GTX 1650 performs just like the current GTX 1050 Ti. Based on the leaked benchmark, the performance difference between the two graphics cards is less than 1 percent. But let's not jump to conclusions just yet. The GTX 1650 could very well be the mobile variant, which would lead to a lower score. Plus, that graphics card could easily be an engineering sample, which would mean it's still susceptible to further tweaking from Nvidia. Also, there's the matter of having compatible drivers for the graphics card to truly perform properly.

Quote:Context matters, in this case, because Nvidia has already told investors that Turing sales were below expectations in the back half of the year. This agreed with publicly available data points from companies like Amazon and Newegg, where the best-selling GPU lists last fall were full of GTX 1080s and 1070s, with far fewer Turing cards making the cut. Nvidia may indeed be earning more money from Turing at launch than it did from Pascal, but the company’s own remarks indicate Turing didn’t meet Nvidia’s initial expectations. Nvidia’s gaming revenue also fell like a rock in Q4 2018 as part of the overall crypto hangover. This doesn’t speak to a strong launch period for its new architecture in terms of unit volumes, even if revenue rose.
...
As for the claim that 90 percent of Nvidia’s market is below the performance of a 1660 Ti? That looks pretty true. Using the Steam Hardware Survey’s numbers (and classifying the 1070 and 1660 Ti as equivalent), there are just four GPUs with any significant market share above the 1660 Ti’s performance: GTX 1070, 1070 Ti, 1080, and 1080 Ti. Add them up, and that’s 10.65 percent. The only other GPU that might beat the 1660 Ti would be the Maxwell-based GTX 980 Ti. (The original GTX 980 was almost exactly equal to the performance of a GTX 1060, so the 980 Ti has a shot at beating the 1660 Ti, though I haven’t done the comparison.) The number of GTX 980 Ti’s in-market is probably pretty small at this point. And since the Steam Hardware Survey isn’t identical to the total GPU market, I think we can call this one pretty straight.

And that’s kind of remarkable in and of itself. The GPUs that drive marketing forward are a far cry from the cards that most people practically purchase. While a full 50 percent of Nvidia’s customer base is using Pascal hardware, the bulk of them bought in at the 1060 level or below. Of course, the flip side to this is that it’ll be a long time indeed before RTX features meaningfully hit the mainstream. If 90 percent of Nvidia’s market is plugged in below the new $280 price point and RTX features start on $350 cards, the best RTX owners can reasonably expect is for 10-15 percent of their fellow gamers to have access to RTX by the time Nvidia’s next generation launches. That’s definitely enough support to drive some adoption, but not enough to build ray tracing into a mainstream feature this generation.

Quote:However, analysis from Digital Foundry today finds that DLSS can actually offer a performance bump of up to 65 percent in Anthem. There is a caveat: Digital Foundry observed a broad range of performance gains with DLSS enabled. Sometimes it was around 20 percent; other times it hit that 65 percent limit. It depends on what's happening in-game and whether DLSS addresses whatever bottleneck is causing the frames to drop. And, of course, the usual clause about individual performance gains being incredibly system- and environment-dependent apply here.
...
The real takeaway is that DLSS could still be hiding its true potential. Early attempts to use the technology might have missed the mark, or at least been limited to a roughly 40 percent improvement to performance, but that will probably change as developers grow familiar with the technology. For people who value performance over everything else, that could be more exciting than the graphical improvements offered by ray tracing.

Quote:The slide shows Acer planning to offer both models with two graphics card options, the GTX 1660 Ti or GTX 1650. As a quick refresher, the GTX 16-series graphics cards are based on the chipmaker's latest Turing architecture but lack RT cores for real-time ray tracing and Tensor cores for AI workloads.

Quote:According to Google Translate, the GTX 1650 will replace the 1050 (no surprises there) but will be significantly faster. MSI reveals some 3DMark figures that claim the GTX 1650 scores 11294 points against the 1050's 7979 and the 1050Ti's 9114. MSI also says the GTX 1650 will provide 60 or more frames per second in games like Assassin's Creed, Fortnite, PUBG, and GTA V. The GTX 1650 will have 4 GB of VRAM, just like the mobile 1050, and its clock speed is 1395 MHz (MSI's promotional material doesn't specify if this is a sustained clock speed, base clock speed, or turbo clock speed).

Quote:Of Nvidia's RTX 20 series, the 2060 gained the most share in March, followed by 2070, then 2080, and finally the 2080Ti. All of the RTX models have individually surpassed the RX Vega series, even the 2080Ti, which finally solidified its lead against Vega in March after tying in February. Despite Vega being much cheaper and offering better value, RTX seems to entice gamers more.
...
On the other hand, the 10-series, of which the 1070Ti is a member, has been shrinking in size. Although the GTX 1660 and 1660Ti are not on this survey, it is likely Nvidia's previous mid-range 10-series GPUs are losing market share to these new GPUs, as well as the slightly more expensive 2060. Overall market share has not changed significantly between Nvidia and AMD, standing at 74.75% and 14.9% respectively. Though RTX is a fast-growing segment for Nvidia, it is likely that users buying RTX previously owned 10-series hardware, offsetting the gains. Meanwhile, AMD is gaining share from gamers looking to upgrade on the cheap.

Quote:After analyzing full, high-res images of NVIDIA's TU106 and TU116 chips, reddit user @Qesa did some analysis on the TPC structure of NVIDIA's Turing chips, and arrived at the conclusion that the difference between NVIDIA's RTX-capable TU106 compared to their RTX-stripped TU116 amounts to a mere 1.95 mm² of additional logic per TPC - a 22% area increase. Of these, 1.25 mm² are reserved for the Tensor logic (which accelerates both DLSS and de-noising on ray-traced workloads), while only 0.7 mm² are being used for the RT cores.

According to the math, this means that a TU102 chip used for the RTX 2080 Ti, which in its full configuration, has a 754 mm² area, could have done with a 684 mm² chip instead. It seems that most of the area increase compared to the Pascal architecture actually comes from increased performance (and size) of caches and larger instruction sets on Turing than from RTX functionality. Not accounting to area density achieved from the transition from 16 nm to 12 nm, a TU106 chip powering an RTX 2060 delivers around the same performance as the GP104 chip powering the GTX 1080 (410 mm² on the TU106 against 314 mm² on GP104), whilst carrying only 75% of the SM count (1920 versus 2560 SMs).

Quote:Nvidia is expected to launch its smallest GPU based on the Turing architecture, the GTX 1650, on April 22. We’ve seen pictures that point to a small form factor reference version sharing a shroud styled like the full-size Founders Edition cards but using a single blower type fan instead. And today, VideoCardz shared what it claimed is the first look at the GTX 1650 from Zotac. It features an axial style fan and does not require any PCIe power.

Quote:A Twitter user going by the handle momomo_us this week discovered results of an unnamed Nvidia graphics card with 7.5GB of onboard memory in the UserBenchmark database. According to the benchmark scores, the mysterious graphics card's performance is right in between that of the Nvidia GeForce RTX 2070 and RTX 2080, which leads us to believe that it could be an RTX 2070 Ti.

On the other hand, it's possible that UserBenchmark failed to pick up the specifications correctly. The unknown device could be an RTX 2070 Ti or it could be something completely different. On top of that, we haven't seen any other evidence of Nvidia releasing an RTX 2070 Ti. As always, take this leak with a good amount of salt.
...
Based on the posted UserBenchmark scores, the RTX 2070 Ti could be, on average, up to 7 percent faster than the RTX 2070 and just 9 percent slower than an RTX 2080. Before drawing a conclusion, it's important to note that the average scores for the RTX 2070 and RTX 2080 consist of over tens of thousands of individual scores, including overclocked models, while the purported RTX 2070 Ti average score is based off nine results. The actual performance gaps between the trio of graphics cards could vary significantly.

From a pricing perspective, an RTX 2070 Ti would be a welcome addition to the Turing RTX 20-series. The RTX 2070 and RTX 2080 debuted at $500 and $700, respectively. So, there is certainly room for a $600 Turing-powered graphics card. Current pricing on the market for custom RTX 2070s range from $470 to $630, while custom RTX 2080s span from $680 to $900. If priced adequately--and real--the RTX 2070 Ti could be one of Nvidia's winning horses.

Quote:NVIDIA today announced that it is extending DXR (DirectX Raytracing) support to several GeForce GTX graphics models beyond its GeForce RTX series. These include the GTX 1660 Ti, GTX 1660, GTX 1080 Ti, GTX 1080, GTX 1070 Ti, GTX 1070, and GTX 1060 6 GB. The GTX 1060 3 GB and lower "Pascal" models don't support DXR, nor do older generations of NVIDIA GPUs. NVIDIA has implemented real-time raytracing on GPUs without specialized components such as RT cores or tensor cores, by essentially implementing the rendering path through shaders, in this case, CUDA cores.

Quote:It’s good to see that owners of older cards (plust the recent non-RTX GTX 1660 and 1660 Ti) will at least be able to try out some of these features and see them with their own eyes. Of course, if you can see ray tracing and like it, but the games don’t run smoothly, that works out in Nvidia’s favor. Because you’ll probably then want to upgrade to an RTX card to achieve better ray tracing results and higher frame rates.

And if you like gaming at high resolutions with ray tracing, you will probably need to upgrade. Note that all of Nvidia’s benchmark results below show Pascal and Turing cards running the existing games and demos at 1440p. And the frame rates for even the GTX 1080 Ti don’t look pretty.

Note that these tests were run at high game and high ray tracing/DXR settings. You’ll be able in many instances to dial down settings and see smoother results. To find out how smooth, we’ll have to spend some quality time dusting off some older Pascal cards and dropping them into our test rig to see what they can do with today’s cutting-edge game features.

Quote:NVIDIA is releasing its most affordable graphics card based on the "Turing" architecture, the GeForce GTX 1650, on the 23rd of April, starting at USD $149. There doesn't appear to be a reference-design (the GTX 1660 series lacked one, too), and so this GPU will be a partner-driven launch. Based on NVIDIA's smallest "Turing" silicon, the 12 nm "TU117," the GTX 1650 will pack 896 CUDA cores and will feature 4 GB of GDDR5 memory across a 128-bit wide memory interface.

Quote:So what does this investigation tell us overall?

• There was no need to test any GTX GPU slower than the Titan X. Two of the three games are already unplayable at 1080p with the Titan.
  
• The only test condition that was remotely usable was Battlefield V at 1080p with Low ray tracing, but even then I would expect a GTX 1080 to barely hit 30 FPS in the most intensive areas and performance to fall away further from there.
  
• Unless you have one of the top-end GTX cards and the ray traced effects aren’t too intensive, Pascal is not fast enough to give gamers a usable ray tracing experience. As more ray traced games are released, Pascal is only going to fall further behind.
  
• In a best case scenario, the Titan X matches the RTX 2060 for ray tracing capabilities, but often falls 30% behind or more, especially when looking at crucial 1% low data. The Titan also delivers less frame rate consistency. Normally (no DXR) the Titan X is at least 25% faster than the RTX 2060 and more, up around RTX 2080 territory.
  

...
Finally we must wonder, why has Nvidia bothered testing and enabling ray tracing on Pascal? It doesn’t run well even on high-end GPUs, it’s unlikely to improve with future games, and it just seems like something people wouldn’t use even if it was available.

We have two theories as to why: The first is for developers. Lets say you have a development studio that invested heavily into Pascal and has loads of cards like this very Titan X in their dev machines. Rather than forcing developers into upgrading to Turing to develop games with ray tracing, allowing Pascal cards to ray trace, albeit slowly, could improve the adoption of ray tracing in games. Developers don’t need 60 or even 30 FPS to test ray tracing in their games, so it could be handy for them. This has a variety of benefits for Nvidia as they are interested in improving adoption to sell more RTX cards.

Second: plain old marketing to incentivize RTX GPU upgrades. In our opinion ray tracing alone isn’t enough to justify an upgrade but this will work for a number of Pascal gamers regardless.

Having features unlocked for more GPU owners is generally a good thing and down the line there might be a ray traced game that actually does run well on Pascal. But like most of our articles on ray tracing so far, we’re still in the very early stages for the technology and it will only become a significant factor in a few generations time.

Quote:The bottom line is that real-time ray tracing is a tremendously expensive operation, even on GPUs with fixed-function RT cores dedicated to its acceleration. Leveraging DXR’s fallback layer to perform those calculations on the CUDA cores hits frame rates significantly harder. But the result isn’t the slide show we might have expected. In fact, in a game like Battlefield V, enthusiasts with high-end Pascal-based cards can enjoy playable performance using the High and Ultra quality presets. The more taxing ray tracing techniques used in SotTR and Metro Exodus certainly add beauty to those titles. Just don’t expect smooth frame rates unless you own one of the faster GeForce RTX models. Cynical though it might be, we have to imagine that’s the exact idea Nvidia is encouraging here.

Quote:Nvidia's GeForce GTX 1650 is slated to launch on April 23, and it seems that the GeForce GTX 1650 Ti might not be far behind. Last week, Asus registered a plethora of custom GeForce GTX 1650 models with the EEC (Eurasian Economic Commission) with a few GeForce GTX 1650 Ti thrown into the mix.
...
According to Asus' EEC filing, the GeForce GTX 1650 Ti features 4GB of memory just like its soon-to-be-released little brother, the GeForce GTX 1650. Asus didn't specify the type and speed of the memory or the size of the memory interface. Looking back at history, the GeForce GTX 1050 and GTX 1050 Ti shared the same memory configuration. If Nvidia continues that trend, it's feasible that the GeForce GTX 1650 Ti might use the same GDDR5 memory running at 2,000MHz (8,000MHz effective) and 128-bit memory bus as the GeForce GTX 1650.