Graphics Cards - Page 134

We all know NVIDIA is launching its new flagship GeForce RTX 3090 Ti graphics card this week at CES 2022, but now EVGA is joining the fun with a tease of its flagship GeForce RTX 3090 Ti KINGPIN graphics card.

EVGA's new custom GeForce RTX 3090 Ti KINGPIN graphics card will reportedly require not one, but two 12-pin PCIe power connectors -- yeah, those PCIe 5.0 PSUs and power connectors -- with 12 normal pins and 4 data paths which are optional. We don't know what manufacturers are calling this yet, but it seems to be compatible with NVIDIA's existing 12-pin PCIe power connector on the RTX 30 series Founders Edition GPUs.

But get this... with dual 12-pin PCIe power connectors under its belt, EVGA's custom GeForce RTX 3090 Ti KINGPIN graphics card will have a theoretical maximum power consumption of up to 1275W. Yes, twelve hundred and seventy five watts. Compared to the EVGA GeForce RTX 3090 KINGPIN which has a 525W TDP, this is a gigantic upgrade of close to 750W with the new EVGA RTX 3090 Ti KINGPIN. Holy balls.

In the hours leading up to CES 2022 we're getting more concrete details about AMD's next-gen Ryzen 9 6980HX processor -- codenamed "Rembrandt".

The chip in the above photo is of the AMD Rembrandt APU, inside of Alienware's upcoming m17 R5 gaming laptop. This is the Ryzen 9 6850HX processor, with an RDNA 2-based GPU next to it in the form of the Radeon RX 6700M which will have 10GB of GDDR6 memory. The m17 R5 gaming laptop will ship with the higher-end AMD Radeon RX 6850M XT graphics card with 12GB of GDDR6 memory.

The new Rembrandt APU arrives with a flagship Ryzen 9 6980X processor that has 8 cores and 16 threads on the refreshed Zen 3+ architecture, and for the first time on the Zen architecture: pushing 5.0GHz. AMD will most likely be pre-binning the chips, with the 45W power target most likely good for the most part -- but it will go over -- by how much, we'll find out soon.

AMD is reportedly gearing up to announce its new Radeon Super Resolution (RSR) technology at CES 2022, which would enable FSR 1.0 in nearly "all" games.

According to VideoCardz, RSR will be based on the FSR 1.0 algorithm and will work simply through the Radeon software drivers. This means that there won't be any need for proprietary super-resolution technology (like FSR, or DLSS) and now with RSR, wouldn't be required in the graphics pipeline.

There's another benefit here: no issues with the UI (user interface) in games, as they're usually rendered at the native resolution and can sometimes look a bit crappy with upscaling technology enabled. RSR won't require game developers to add support, as it'll work independently through the Radeon drivers.

Innosilicon has officially launched its new graphics cards based on its in-house Fantasy One GPU, with 4 new graphics cards based on the Fantasy One GPU launched -- including a multi-GPU design -- in collaboration with GPU designer Imagination Technologies.

We don't have any concrete performance numbers, which we shouldn't expect anything less from the Chinese manufacturer, with a demo posted on social media with a video that doesn't include an FPS counter, or much at all. The new Fantasy One GPU-powered Type-A graphics card -- yeah, that's the name of the card and not USB Type-A -- the Type-A graphics card.

Inside, a single Fantasy One GPU has up to 16GB of GDDR6X memory on a 128-bit memory interface with both 4GB and 8GB offerings. The 16GB of GDDR6X memory is clocked at 19Gbps with 304GB/sec of memory bandwidth -- while display connectivity will include DisplayPort 1.4, HDMI 2.1, and some old-school VGA.

NVIDIA is reportedly coughing up big money to TSMC (Taiwan Semiconductor Manufacturing Company) to ensure that it secures plenty of next-gen 5nm wafers for its upcoming Ada Lovelace GPUs that will power the next-gen GeForce RTX 40 series graphics cards.

But in new rumors from MyDrivers, NVIDIA is paying much more than usual to get 5nm wafer from TSMC with the company reportedly paying TSMC a large $1.64 billion in Q3 2021... but that increases to $1.79 billion in Q1 2022 as they ramp up to the Ada Lovelace GPU being made and printed onto next-gen GeForce RTX 40 series GPUs.

A larger long-term deal has also reportedly been struck, which would see NVIDIA coughing up $6.9 billion to TSMC in order to secure next-gen 5nm wafers for Ada Lovelace. We've heard before that NVIDIA's next-gen GeForce RTX 40 series GPUs would be made by TSMC -- and not Samsung -- and its upcoming 5nm node.

AMD is set to launch its new Radeon RX 6500 XT graphics card in the coming weeks, powered by the new cut-down Navi 24 GPU on the fresh 6nm node by TSMC.

The smallest member of the RDNA 2 family of graphics cards is close, with the budget desktop gaming GPU -- the Navi 24 GPU itself with a die size of an estimated 141mm2 -- being smaller than the RDNA-based Navi 14 with 158mm2. We should see the new Navi 24 GPU with half the GPU clusters of Navi 23 (which has 16).

Each of the GPU clusters have 64 Stream Processors, with the full GPU offering 1024 shading units -- AMD has also reduced its Infinity Cache on the Navi 24 GPU -- down to 16MB Infinity Cache on a 64-bit wide memory controller. This should be the last GPU in the RDNA 2 family of chips, however, it is the first 6nm GPU that AMD will be launching (when it launches). AMD will also be most likely launching its mobile-focused "Rembrandt" APU on 6nm, too.

NVIDIA will be detailing not one but two new mobile GPUs at CES 2022 next week, with the flagship GeForce RTX 3080 Ti Laptop GPU receiving some more (leaked) details before its announcement.

The new GeForce RTX 3080 Ti Laptop GPU will reportedly have 16GB of GDDR6 memory clocked at 16Gbps, which is up from the RTX 3080 Laptop GPU variants and their 8GB/16GB of GDDR6 -- but only at 14Gbps. NVIDIA is expected to ramp up the TGP on the GeForce RTX 3080 Ti Laptop GPU to 175W (10W more than the RTX 3080 Laptop GPU).

We might see AIBs making custom gaming laptop solutions with NVIDIA's new flagship GeForce RTX 3080 Ti Laptop GPU inside, with a few already teased ahead of CES 2022. But there's one thing to note here: the maximum TGP isn't some magic 175W ceiling -- but rather, laptop manufacturers could have the ability to boost past that -- we'll know more on that soon.

NVIDIA is ready to unleash its new flagship GeForce RTX 3090 Ti graphics card, and AIB partners are close behind with some slick new custom GeForce RTX 3090 Ti graphics cards which should knock our socks off.

Our friends at MSI have their new custom GeForce RTX 3090 Ti SUPRIM X in the silicon oven right now, with VideoCardz uploading some images and an NDA document from an anonymous source. The NDA itself covers the name of the card -- MSI GeForce RTX 3090 Ti SUPRIM X 24G -- while the NDA embargo lift is on January 27, 2022 @ 6AM PT, 3PM CET, 10PM CST.

MSI isn't the only one preparing a new custom GeForce RTX 3090 Ti graphics card, with ASUS also working on its custom GeForce RTX 3090 Ti TUF Gaming which was recently teased.

Intel is preparing for a big fight in the GPU market in 2022, but the real fun -- and the gloves come off -- in 2023 when the company launches its next-next-gen DG3 "Battlemage" GPU.

In some fresh new rumors from Tom @ Moore's Law is Dead, Intel's next-gen DG3 Battlemage series of GPUs codenamed "Elasti" are going to be monsters. MLID's sources have said that Intel's new DG3 Battlemage GPU will be taking the good fight to both AMD and NVIDIA in the high-end flagship graphics card market, but we've got a year+ to wait for that.

We only heard about Intel's codename "Elasti" GPU a few days ago, but between now and then we'll have the launch of the Intel Xe-HPG "Arc Alchemist" GPUs in 2022, which will be followed by the Intel Xe2-HPG "Arc Battlemage" GPU in 2023 before we see the likes of Intel Xe3-HPG "Arc Celestial" in 2024+ while Intel Xe Next "Arc Druid" and Intel Xe Next Next "Arc Elasti" will probably launch in 2025 and 2026, respectively.

COLORFUL has just introduced two new custom GeForce RTX 2060 12GB graphics cards, with the introduction of the new GeForce RTX 2060 12GB iGame Ultra White and BattleAx Deluxe graphics cards.

The company is using its Ampere-ready GeForce RTX 3060 cooling solutions for the new Turing-based GeForce RTX 2060 12GB graphics cards in the new Ultra White and BattleAx Deluxe offerings. The new COLORFUL RTX 2060 12GB iGame Ultra OC White sports a premium design and factory overclocking, as well as a custom OC button on the I/O bracket.

The original GeForce RTX 2060 6GB version of the card had 2 x DisplayPort 1.4, HDMI, DVI, and USB Type-C, while the new COLORFUL RTX 2060 12GB graphic cards both have just a single DisplayPort 1.4, HDMI 2.1, and DVI connector. DVI, in 2022, alright then. We have GPU clocks of up to 1665MHz on the Ultra OC variant, while there's 1650MHz GPU clocks on the BattleAx Deluxe.

Chinese GPU manufacturer Jing Jiawei has just announced that their new JM9 series processor is for sale, and while it doesn't have record-breaking performance, we do have a new player in the market.

Jing Jiawei will meet the premium requirements required for displays, AI computing, media processing, games, geographic information systems, CAD-based designing, and virtualization. The new JIM9 series processor is aiming at entry-level markets, with the company prompting an institutional survey that showed the JM9 series GPU was in contact with specifically targeted customers, and are prepared to start debugging and optimization in 2022.

8GB of memory at up to 128GB/sec of memory bandwidth, which should see the GPU on a 128-bit memory bus. This means the GPU is most likely coupled with older GDDR6 memory, but it could be GDDR6 -- I guess we'll find out soon enough. Maybe.

Optical computing startup Lightelligence has developed a new GPU that outstrips the very best GPUs on the planet -- at least in the most challenging mathematical problems -- with the introduction of the Photonic Arithmetic Computing Engine (PACE).

Lightelligence's new Photonic Arithmetic Computing Engine (PACE) thrashes NVIDIA's third-fastest Turing-based GPU in the GeForce RTX 3080 by nearly 100x in the NP-complete class of problems. The new PACE accelerator under the Ising model -- an example of a thermodynamic system that's used for understanding phase transitions -- with record-breaking results.

The new PACE accelerator when compared to the GeForce RTX 3080 is around 100x faster, with Lightelligence using 12,000 optical devices integrated onto a circuit, clocked at 1GHz. Up against the purpose-built Toshiba's simulated bifurcation machine based on FPGAs, Lightelligence's new PACE accelerator still beats that system by 25x -- just as impressive.

The purported new Intel Arc Alchemist graphics card has been spotted again, this time running on the Ashes of the Singularity benchmark. This is interesting, as it might have something to do with what we'll see from Intel at CES 2022.

The system was powered by Intel's new flagship Core i9-12900K processor, with what would be an engineering sample of the Xe-HPG DG2 GPU aka Intel Arc Alchemist a320. Intel's new Arc Alchemist graphics card was benchmarked at 1080p on the Medium preset, scoring 12,500 points and an average of 126.9FPS.

Intel is going to have a beefy first GPU -- it won't beat NVIDIA or AMD, but it'll do quite well -- especially the flagship Xe-HPG 512 EU version. This GPU should have 16GB of GDDR6 memory @ 16-18Gbps on a 256-bit memory bus, with 225W TDP on the desktop (and 120-150W on the laptop GPU).

NVIDIA is expected to launch its new flagship GeForce RTX 3090 Ti graphics card at CES 2022 in just a couple of weeks time, and we're already seeing custom RTX 3090 Ti graphics card in the wild.

The new custom ASUS GeForce RTX 3090 Ti TUF Gaming graphics card has been spotted, with the folks at ITHome teased our first look at the ASUS GeForce RTX 3090 Ti TUF Gaming graphics card -- well, at least the box -- and on the box we can see that it seems there's no PCIe 5.0 support.

In previous rumors we heard PCIe 5.0 support would be featured on NVIDIA's new beefed-up GeForce RTX 3090 Ti, which will have the same 24GB of GDDR6X memory that the GeForce RTX 3090 has -- but it will be new single-sided GDDR6X, which should radically reduce the GDDR6X temperatures on the backplate of the RTX 3090 Ti.

SAPPHIRE has just introduced its new NITRO+ S360-A and S240-A AIO CPU coolers, which will work with the latest AMD AM4 and Intel LGA1700 socket processors.

The new SAPPHIRE NITRO+ AIO CPU coolers use the company's in-house NITRO+ Hybrid Fan Blade, which uses a combination of a traditional axial fan (which lacks the pressure to push air down onto the components) and a blower fan (which is strong, but is noisy when the fan speed is cranked) and then improved the downward air pressure of the axial fan design... while keeping the noise low.

This should result in fantastic thermal performance for cooling the very best AMD Ryzen 5000 series or Intel's new Alder Lake CPUs, all while keeping things quiet -- and man, do they look good. There's some great design work here by SAPPHIRE, as well as some gorgeous RGB lighting that will look sweet in the right system -- especially if you mixed it with SAPPHIRE's own TOXIC Radeon RX 6900 XT Limited Edition graphics card. Oh man, would that look good.

NVIDIA is preparing not just one but two new GeForce RTX 3050 desktop GPUs in 8GB, and 4GB variants according to the latest leaks.

The new GeForce RTX 3050 will launch in two versions, the first is with the GA106-150 GPU with 2560 CUDA cores and 8GB of GDDR6 while the other is will have the GA106-140 GPU with 2304 CUDA cores and 4GB of GDDR6 memory. The other rumors of the GPU inside of the RTX 3050 had it pegged with 3072 CUDA cores, so these new rumors are very different.

Not only do we have a GPU with a different CUDA core count, but multiple versions of the RTX 3050: one with 8GB, one with 4GB. We should expect NVIDIA to unveil in early 2022, where we should see the company also unveil and release its powerful new GeForce RTX 3090 Ti graphics card, as well as the laptop-focused GeForce RTX 3080 Ti and GeForce RTX 3070 Ti Laptop GPUs at CES 2022 in just a few weeks time.

AIDA64's latest version now supports NVIDIA's unreleased GeForce RTX 3080 Ti Laptop GPU and Intel's new 13th Gen "Raptor Lake" CPUs coming out in 2022.

The new AIDA64 Extreme 6.60 software is used for benchmarking, professional diagnoses, and more for professionals, engineers, benchmarkers, gamers, and more. The new AIDA64 release has optimized benchmarks for Intel's new 12th Gen "Alder Lake" and next-gen 13th Gen "Raptor Lake" CPUs.

There's also improvements for DDR5 memory modules and XMP 3.0 memory profiles, as well as support for the very latest AMD and Intel chipsets. On the GPU side of things, we have support for NVIDIA's unreleased GeForce RTX 3080 Ti Laptop GPU, and the new desktop release of the GeForce RTX 2060 12GB.

NVIDIA is reportedly preparing a beefed-up GeForce RTX 3080 12GB graphics card, a new model with 2GB more GDDR6X memory, more CUDA cores, and more performance across the board.

The new GeForce RTX 3080 12GB graphics card will have its 12GB of GDDR6X memory on a wider 384-bit memory bus, which will see up to 912GB/sec of memory bandwidth -- quite the increase over the 760GB/sec memory bandwidth on 10GB of GDDR6X memory on the regular GeForce RTX 3080.

NVIDIA will be using the new GA102-220 GPU for the GeForce RTX 3080 12GB, which packs more CUDA cores than the GA102-200 inside of the RTX 3080. The new GA102-220 GPU has 8960 CUDA cores (256 more CUDA cores than the RTX 3080 GPU) with 2 more RT cores, and 8 more Tensor Cores/TMUs.

Intel has teased a bunch of its unreleased CPUs and GPUs in its new drivers for the 11th Gen NUC, with codenames for GPUs that were previously leaked, now confirmed.

Intel's next-gen DG3 "Elasti" GPU has been confirmed, as well as the Arc Alchemist DG2 graphics cards and CPU codenames for next-gen Intel Meteor Lake, Arrow Lake, and Lunar Lake CPUs. Inside, leaker KOMACHI_ENSAKA noticed mentions of "ELG" which later turned into what we know as the Intel DG3 "Elasti" GPU.

The driver number of Intel DG3 "Elasti" ended in "9999" which references this driver is a "testing driver" which isn't available to the public, but now is available on Intel's official website. As for "Elasti", that is the codename to Intel's next-next-gen DG3 discrete GPU, but the driver also supports multiple Intel next-gen CPUs.

Inside of Samsung's new product catalog we now find that the South Korean giant is sampling 16Gb (2GB) memory modules based on GDDR6 technology with a bandwidth of 20Gbps (K4ZAF325BC-SC20) and 24Gbps (K4ZAF325BC-SC24) to its partners.

The new modules are on Samsung's product catalog, but we don't know what products Samsung's new GDDR6 memory will be used on first. AMD's new Radeon RX 7900 XT could be the first card to feature Samsung's new GDDR6 memory clocked at the impressive 24Gbps, up from the 24GB of GDDR6X memory clocked at 19.5Gbps on the GeForce RTX 3090 and the 21Gbps that you can overclock to 21Gbps.

The refreshed GeForce RTX 3090 Ti is meant to have the same 24GB of GDDR6X, but at a much faster 21Gbps -- the same 21Gbps that the double-sided GDDR6X on the RTX 3090 can do with an easy OC. But remember, that's GDDR6X -- the X is a big difference as Micron makes it, and it costs more than GDDR6.