This generation's GPU release is shaping up to be one of the most interesting in late years; for some good reasons, and bad reasons alike. We've heard - keep in mind, not seen - NVIDIA back down from multiple graphics card releases (the double VRAM versions of RTX 3070 and 3080 come to mind); postponing the RTX 3070 until after they have gleaned exactly what AMD will be offering with their RX 6000 series; preparing to launch an RTX 3060 Ti with no announcement whatsoever and before the RTX 3060 is ever launched; and now, apparently, the company is readying a response to AMD's as-of-yet-unannounced RX 6000 series in the form of the RTX 3080 Ti.

Recent performance leaks have placed an unclear AMD GPU (and apparently, not even the fastest Big Navi chip at that) at the same performance level as NVIDIA's RTX 3080, which is a tremendous increase in performance for the red team, coming from years of only being able to effectively compete in the midrange offerings. Now, Kopite7kimi, a known leaker with a proven track record, has claimed that NVIDIA is already prepping a new GA102-based graphics card, sitting in performance between the RTX 3080 and RTX 3090. He lists identifies the GPU as GA102-250-A1, 9984FP32, 384bits GD6X - let's call it the RTX 3080 Ti. If those details are correct, this is yet another product demanding the same 628 mm² GA102 GPU be available for it - in a scenario with inadequate availability of the RTX 3080, 3090, and likely 3070 Ti graphics cards already, should that later one actually materialize. The memory bus on this prospective RTX 3080 Ti is apparently inheriting the same design as the RTX 3090, with a 384-bit solution (compared to the RTX 3080's 320 bit), and likely 12 GB of GDDR6X memory.

GIGABYTE backed up its late September launch of the RTX 3080-based VISION OC graphics card targeted at creators, with one based on the GeForce RTX 3090 (model: GV-N3090VISION OC-24GD), a GPU that offers greater dividends to creators thanks to its 24 GB video memory. GIGABYTE's VISION brand of graphics cards and motherboards are targeted at creators, and the RTX 3090 VISION OC, when paired with NVIDIA's GeForce Studio drivers, provides a formidable solution halfway between the gaming and professional-visualization market segments.

The GIGABYTE RTX 3090 VISION OC comes with the same board design as the RTX 3080 VISION OC, but with the addition of the NVLink interface for explicit multi-GPU. The card comes with a mild factory-OC which sees the GPU boost up to 1755 MHz (vs. 1695 MHz reference), while the memory is left untouched at 19.5 Gbps (GDDR6X-effective), for 940 GB/s of memory bandwidth. Display interfaces include three DisplayPort 1.4a and two HDMI 2.1 connectors. The card draws power from two 8-pin PCIe power connectors. It uses a triple-slot, triple-fan cooling solution with the VISION design scheme. The company didn't reveal pricing.

Possibly unsure of the GeForce RTX 3070 tackling AMD's Radeon RX 6000 series parts, NVIDIA is designing a new RTX 30-series SKU positioned between the RTX 3070 and RTX 3080. This is not a 16 GB variant of the RTX 3070, but rather a new SKU based on the 8 nm "GA102" silicon, according to a reliable source with NVIDIA leaks, kopite7kimi. The SKU is based on the GA102 with the ASIC code "GA102-150-KD-A1." The silicon is configured with 7,424 CUDA cores across 58 streaming multiprocessors (29 TPCs), 232 tensor cores, 232 TMUs, 58 RT cores, and an unknown number of ROPs. According to kopite7kimi, the card is configured with a 320-bit wide memory interface, although it's not known if this is conventional GDDR6, like the RTX 3070 has, or faster GDDR6X, like that on the RTX 3080.

NVIDIA recently "cancelled" a future 16 GB variant of the RTX 3070, and 20 GB variant of the RTX 3080, which is possibly the company calibrating its response to the Radeon RX 6000 series. We theorize that doubling in memory amounts may not have hit the desired cost-performance targets; and the company probably believes the competitive outlook of the RTX 3080 10 GB is secure. This explains the need for a SKU with performance halfway between that of the RTX 3070 and RTX 3080. As for pricing, with the RTX 3070 positioned at $500 and the RTX 3080 at $700, the new SKU could be priced somewhere in between. AMD's RDNA2-based Radeon RX 6000 series GPUs are expected to feature DirectX 12 Ultimate logo compliance, meaning that there is a level playing ground between AMD and NVIDIA in the performance segment.

Fresh reports floating in the rumor mill's circulatory system claim that NVIDIA backtracked on its plans to launch higher VRAM capacity versions of their RTX 3080 and the (in the meantime, delayed) RTX 3070. These cards launched with 10 GB VRAM for the RTX 3080 and 8 GB VRAM for the RTX 3070, with reports circulating as early as their announcement that there would be double-capacity versions hitting the market just a few months later - specifically, in December of this year. Videocardz, however, claims that these long-rumored 20 GB and 16 GB SKUs have now been canceled by NVIDIA, who sent this news to its AIB partners - and the usage of canceled, not postponed, is perfunctory.

For cards theoretically shipping come December, this is indeed a small advance notice, but it might be enough for AIB partners to feed all their GA102-200 (RTX 3080) and GA104-400 (RTX 3070) silicon towards the already - if not readily - available models. This report, Videocardz claims, has been confirmed by two of their sources, and comes at the exact same day specifications for AMD's RX 6000 series leaked. It's likely NVIDIA already had knowledge of its competition's designs and performance targets, however, so this could be seen as nothing more than a coincidence. One of the publications' sources claims GDDR6X yields might be the cause for the cancellation, but this doesn't help explain why the alleged RTX 3070 16 GB card (with its GDDR6 chips) was also canceled. Remember: these are rumors on cards that were never announced by NVIDIA themselves, so take these with the appropriate salt-mine level of skepticism.

Alphacool presents the Eisblock ES Acetal GPX-N RTX 3080/3090 GPU watercooling block The block is specially developed for the server sector. Nevertheless, it also fits into normal desktop computers or small form factor cases. Of course, it is also available to our end customers. A special feature of this block is the terminal with the G1/4" ports is located at the rear end of the cooler. This arrangement saves space and makes it much easier to lay and connect the tubing, especially in server cases.

Alphacool naturally only uses copper in its water coolers. In the Eisblock ES Acetal GPX-N RTX 3080/3090 the copper is nickel-plated. Compared to the previous models, Alphacool has improved the type of nickel plating again, which significantly increased the acid resistance.

Let's skip ahead of any "Can it run Crysis" introductions for this news piece, and instead state it as it is: Crysis 3 can absolutely run when installed directly on a graphics card's memory subsystem. In this case, an RTX 3090 and its gargantuan 24 GB of GDDR6X memory where the playground for such an experiment. Using the "VRAM Drive" application, distributed in an open-source manner via the GitHub platform, one can allocate part of their GPU's VRAM and use it as if it was just another system drive. After doing so, user Strife212 (as per her Twitter handle) then went on to install Crysis 3 on 15 GB of the allocated VRAM. The rest of the card's 9 GB were then available to actually load in graphical assets for the game, and VRAM consumption (of both the installed game and its running assets) barely crossed the 20 GB total VRAM utilization.

As you might expect, graphics memory is one of the fastest memory subsystems on your PC, being even faster (in pure performance terms) than system RAM. Loading up of game levels and asset streaming from VRAM "disk-sequestered" pools to free VRAM pools was obviously much faster than usual, even more than the speeds achieved by today's NVMe drives. Crysis 3 in this configuration was shown to run by as many as 75 FPS in 4K resolution, with the High preset settings. A proof of concept more than anything - but users with a relatively powerful (or memory-capable) graphics card can perhaps look at this exotic solution as a compromise of sorts, should they not have any fast storage options, and provided the game install size is relatively small.

[Update, September 29, 2020: We now have a good first look at the card courtesy the man himself. It confirms what we already knew, with a 360 mm AIO cooler and a flip-up OLED screen paired to the flagship offering from EVGA this generation.]

It's only been moments after the RTX 3090's release, but professional overclockers are already unleashing the power available on NVIDIA's GA-102 chip by resorting to exotic cooling techniques. Renowned overclocker Vince "K|NGP|N" Lucido, who works in close proximity with NVIDIA AIB EVGA, tamed the RTX 3090's temperature by resorting to liquid nitrogen. This, alongside tweaks to Vcore (1069 mV) allowed the card to reach a startling 2.58 GHz core clock (a staggering 52.2% increase over NVIDIA's base clock), and 10.750 MHz (21.5 Gbps) memory clocks on the GDDR6X subsystem, which in itself is a 10.3% increase over reference clocks.

The 16.673 3D Mark Port Royal score was achieved with a fully custom design made by EVGA with Vince Lucidos' input. This über 3090 ultimately delivered a performance increase of around 30% more than the stock RTX 3090 would be able to, which isn't a bad equilibrium between the core and memory clock increases. Some might say this is the performance delta one would expect between the 3090 and the 3080 (an overall 40% performance increase, considering the 3090 is already an average of 10% faster than the 3080 at stock clocks).

NVIDIA themselves have shared performance slides for their imminent RTX 3090 graphics card, the new halo product that's been marketed as the new Titan. Previous-gen Titans have achieved extremely meager performance uplifts compared to NVIDIA's top-of-the-line cards (see RTX 2080 Ti vs RTX Titan, an average of 8% performance difference in favor of the Titan. According to the company, users should expect a slightly higher performance uplift this time around, though the 10-15% higher performance in 4K still seems meager - in pure price/performance terms - for the average consumer.

The average consumer who isn't the main focus for this graphics card and its gargantuan 24 GB of GDDR6X memory, anyway - this is more aimed at the semi-professional or professional crowds working with specialized software, whether it be in rendering or AI-based workloads. The RTX 3090 is thus not so much a product for the discerning computer enthusiast, but more of a halo product for gamers, and a crucial product for professionals and academics.

An alleged event by GALAX targeted at distributors in China revealed up to three upcoming SKUs in NVIDIA's RTX 30-series. This comes as yet another confirmation from a major NVIDIA AIC partner about the 20 GB variant of the GeForce RTX 3080. The RTX 3080 originally launched with 10 GB memory earlier this month, and it is widely expected that NVIDIA fills the price-performance gap between this $700 SKU and its $1,500 sibling. The RTX 3080 uses twenty 8 Gbit GDDR6X memory chips (two chips per 32-bit data-path), much like how the RTX 3090 achieves its 24 GB memory amount.

Elsewhere we see GALAX mention the RTX 3060, a performance-segment SKU positioned under the RTX 3070. You'll notice that the product-stack graph by GALAX suggests performance comparisons to previous-generation SKUs. The RTX 3080 and RTX 3090 are faster than everything from the previous generation, while the RTX 3070, which is coming next month, is shown trading blows with both the RTX 2080 Ti and the RTX 2080 Super. In this same graph, the RTX 3060 is shown matching up to the RTX 2080 (non-Super), a card NVIDIA originally launched at $700.

When NVIDIA announced its next-generation GeForce RTX 3080 and 3090 Ampere GPUs, it specified that the memory found in the new GPUs will be Micron's GDDR6X variant with 19 Gbps speed. However, being that there are faster GDDR6X modules already available in a 21 Gbps variant, everyone was left wondering why NVIDIA didn't just use the faster memory from Micron. That is exactly what Igor's Lab, a technology website, has been wondering as well. They have decided to conduct testing with an infrared camera that measures the heat produced. To check out the full testing setup and how they tested everything, you can go here and read it, including watching the video embedded.

Micron chips like GDDR5, GDDR5X, and GDDR6 are rated for the maximum junction temperature (TJ Max) of 100 degrees Celsius. It is recommended that these chips should run anywhere from 0C to 95C for the best results. However, when it comes to the new GDDR6X modules found in the new graphics cards, they are not yet any official specifications available to the public. Igor's Lab estimates that they can reach 120C before they become damaged, meaning that TJ Max should be 110C or 105C. When measuring the temperature of GDDR6X modules, Igor found out that the hottest chip ran at 104C, meaning that the chips are running pretty close to the TJ Max they are (supposedly) specified. It is NVIDIA's PCB design decisions that are leading up to this, as the hottest chips are running next to voltage regulators, which can get pretty hot on their own.

Alphacool presents the Eisblock Aurora Plexi GPX-N cooler for reference design NVIDIA RTX 3080 and 3090 graphics cards. The cooler offers outstanding cooling performance thanks to its full cover design. The new backplate, which is delivered together with the cooler, also contributes to this. This stabilizes the graphics card and ensures an even contact along the whole cooler. The water block cold plate is made of nickel-plated copper. The cooler covers all relevant components such as voltage converters and the graphics memory.

The Alphacool Eisblock GPX-N Aurora Plexi has an LED strip across the entire width with addressable 5 V digital RGB LEDs. These illuminate the entire cooler homogeneously and ensures an outstanding ambiance in the PC housing.

Lenovo has allegedly listed an unreleased NVIDIA GeForce "Ampere" graphics card as an option for its Legion T7 gaming desktop. The specs sheet speaks of a "GeForce RTX 3070 Ti" with "16 GB of GDDR6" memory, in what may be the first of many future product stack refreshes NVIDIA has planned. No other specs of the RTX 3070 Ti have been mentioned, but the doubling in memory size foreshadows a similar treatment for the RTX 3080. Currently, the RTX 3080 is being released with 10 GB of GDDR6X. The flagship RTX 3090 has 24 GB of it. There is a vast gorge in the new product stack, between the $700 RTX 3080 10 GB, and the $1,500 RTX 3090 24 GB, which is ripe for a possible "RTX 3080 Ti" with 20 GB of memory, among other features.

Micron Technology, Inc., today announced the world's fastest discrete graphics memory solution, GDDR6X, the first to power system bandwidth up to 1 terabyte per second (TB/s). Working with visual computing technology leader NVIDIA, Micron debuted GDDR6X in the new NVIDIA GeForce RTX 3090 and GeForce RTX 3080 graphics processing units (GPUs), which are tailored to support the fast speeds that immersive, high-performance gaming applications demand.

"Our multilevel signaling innovation in GDDR6X has shattered conventional bandwidth limits, clocking record-breaking speeds," said Tom Eby, senior vice president and general manager of the Compute & Networking Business Unit at Micron. "Unlike traditional memory, GDDR6X has unparalleled data rates that can keep pace with gaming innovation and data-hungry applications - setting a new standard for graphics memory."

PNY today announced the expansion of its NVIDIA GeForce RTX GPU family by adding the new PNY XLR8 Gaming GeForce RTX 30 Series. The PNY XLR8 Gaming GeForce RTX 30 Series line of GPUs will be available in three different configurations: RTX 3090, RTX 3080 and RTX 3070, all powered by the all-new NVIDIA Ampere architecture, providing enhanced gaming realism, powerful real time ray tracing, extreme overclocking capabilities, and cutting-edge features that will power-up your performance and accelerate the way you game.

The new NVIDIA GeForce RTX 30 Series GPUs, the 2nd generation of RTX, features new RT Cores, Tensor Cores and streaming multiprocessors, bringing stunning visuals, amazingly fast frame rates, and AI acceleration to games and creative applications. Powered by the NVIDIA Ampere architecture, which delivers increases of up to 1.9X performance-per-watt over the previous generation, the RTX 30 Series effortlessly powers graphics experiences at all resolutions, even up to 8K at the top end. The GeForce RTX 3090, 3080, and 3070 represent the greatest GPU generational leap in the history of NVIDIA.

Here are some of the clearest pictures of an NVIDIA GeForce RTX 3090 "Ampere" Founders Edition PCB, albeit bare (components not yet placed). The PCB goes in this form to another production line, where a PCB placer machine with reels of components places them along both sides. We can still make out quite a few things. To beign with, there are six high current traces near where the 12-pin Molex MicroFit power connector goes in.

NVIDIA seems to have utilized both sides of the PCB effectively. The card features traces for a mammoth 20-phase VRM. The obverse side has the chokes and DrMOS, the reverse side has the capacitors. Twelve memory pads for the 384-bit wide GDDR6X memory interface surround the GPU pad. There are pads for four display output connectors, which could be a mixture of DisplayPort, HDMI, or even USB-C VirtualLink. The PCB still has its spacers on, which will be cut out after component placement, after which the PCB heads for testing and product assembly. NVIDIA is expected to announce the GeForce RTX 3090 Ampere at an online event later today.

A mega dump of the Gainward GeForce RTX 3090 Phoenix GS and RTX 3080 Phoenix GS reveal not only the common board design of the two cards, but also the final specs of the RTX 3080 and RTX 3090. The RTX 3090 features 5,248 CUDA cores, and 24 GB of 19.5 Gbps GDDR6X memory across a 384-bit memory bus, which belts out 936 GB/s of memory bandwidth. The Gainward Phoenix GS runs the RTX 3090 at 1725 MHz boost frequency. The RTX 3080, on the other hand, features 4,352 CUDA cores, and 10 GB of 19 Gbps GDDR6X memory across a 320-bit memory bus, with 760 GB/s memory bandwidth. Gainward is running the RTX 3080 at 1740 MHz on the Phoenix GS.

What's interesting is the board power figures put out by Gainward. The RTX 3090 typical board power (at least for the Phoneix GS), is rated at 350 W, while that of the RTX 3080 is rated at 320 W. These explain why we're seeing custom-design RTX 3090 cards with either three 8-pin PCIe power connectors, or in case of the Founders Edition card, the 12-pin connector that's capable of 600 W power delivery. Many of the custom-design RTX 3080 cards we've come across have two 8-pin PCIe inputs. The slides also list out "2nd generation RTX technology," and "3rd gen tensor cores." Gainward's board features a meaty triple-slot, triple-fan cooling solution that has RGB LED illumination. We predict Palit's cards to look very similar to these (with different cooler shroud designs).Update 06:09 UTC: More pics follow, courtesy harukaze5719.

Just ahead of the September launch, specifications of NVIDIA's upcoming RTX Ampere lineup have been leaked by industry sources over at VideoCardz. According to the website, three alleged GeForce SKUs are being launched in September - RTX 3090, RTX 3080, and RTX 3070. The new lineup features major improvements: 2nd generation ray-tracing cores and 3rd generation tensor cores made for AI and ML. When it comes to connectivity and I/O, the new cards use the PCIe 4.0 interface and have support for the latest display outputs like HDMI 2.1 and DisplayPort 1.4a.

The GeForce RTX 3090 comes with 24 GB of GDDR6X memory running on a 384-bit bus at 19.5 Gbps. This gives a memory bandwidth capacity of 936 GB/s. The card features the GA102-300 GPU with 5,248 CUDA cores running at 1695 MHz, and is rated for 350 W TGP (board power). While the Founders Edition cards will use NVIDIA's new 12-pin power connector, non-Founders Edition cards, from board partners like ASUS, MSI and Gigabyte, will be powered by two 8-pin connectors. Next up is specs for the GeForce RTX 3080, a GA102-200 based card that has 4,352 CUDA cores running at 1710 MHz, paired with 10 GB of GDDR6X memory running at 19 Gbps. The memory is connected with a 320-bit bus that achieves 760 GB/s bandwidth. The board is rated at 320 W and the card is designed to be powered by dual 8-pin connectors. And finally, there is the GeForce RTX 3070, which is built around the GA104-300 GPU with a yet unknown number of CUDA cores. We only know that it has the older non-X GDDR6 memory that runs at 16 Gbps speed on a 256-bit bus. The GPUs are supposedly manufactured on TSMC's 7 nm process, possibly the EUV variant.

NVIDIA's second fastest "Ampere" graphics card to launch this year, the GeForce RTX 3080, surfaced on the Userbenchmark database. Hardware Leaks (aka @_rogame) fished out several juicy details about the card that will be positioned right below the flagship RTX 3090 (RTX 2080 Ti successor) that's been in the news lately. The RTX 3080 succeeds the RTX 2080. On the Userbenchmark database, the purported RTX 3080 is shown bearing a device ID "10DE 2206." Among its readable specs leaked are a GPU frequency of up to 2.10 GHz, possibly frequency capped just like "Turing," and 10 GB of GDDR6X memory across a 320-bit wide memory interface, and a memory clock speed of 19 Gbps (GDDR6X effective), which works out to 760 GB/s of memory bandwidth.

Micron have spilled the beans on at least some specifications for NVIDIA's next-gen Ampere graphics cards. In a new tech brief posted by the company earlier this week, hidden away behind Micron's market outlook, strategy and positioning, lie some secrets NVIDIA might not be too keen to see divulged before their #theultimatecountdown event.

Under a comparison on ultra bandwidth solutions, segregated into the GDDR6X column, Micron lists a next-gen NVIDIA card under the "RTX 3090" product name. According to the spec sheet, this card features a total memory capacity of 12 GB GDDR6X, achieved through 12 memory chips with a 384-bit wide memory bus. As we saw today, only 11 of these seem to be populated on the RTX 3090, which, when paired with specifications for the GDDR6X memory chips being capable of 19-21 Gbps speeds, brings total memory subsystem bandwidth towards the 912 - 1008 GB/s range (using 12 chips; 11 chips results in 836 GB/s minimum). It's possible the RTX 3090 product name isn't an official NVIDIA product, but rather a Micron-guessed possibility, so don't look at it as factual representation of an upcoming graphics card. One other interesting aspect from the tech brief is that Micron expects their GDDR6X technology to enable 16 Gb (or 2 GB) density chips with 24 Gbps bandwidth, as early as 2021. You can read over the tech brief - which mentions NVIDIA by name as a development partner for GDDR6X - by following the source link and clicking on the "The Demand for Ultra-Bandwidth Solutions" document.