Graphics Cards - Page 86

Released over a decade ago, Kepler-based GeForce GTX graphics cards are what you'd consider pretty old when stacked up against the Ampere and Ada Lovelace generation of today. NVIDIA's Kepler architecture for graphics cards debuted in 2012, with the successor to Fermi all about improving energy efficiency alongside performance.

The new Streaming Multiprocessor Architecture, or SMX, was a major milestone in developing GeForce hardware for PC gaming. Kepler GPUs were also the first to introduce a dynamic GPU Boost.

Found inside the GeForce GTX 600 and GTX 700 Series of GPUs, the GK110-400-A1 variant of the high-end 28nm Kepler chip was also inside the very first GeForce GTX TITAN that launched in 2013.

The latest on the NVIDIA RTX 4090 GPU ban over in China is that it's going ahead as planned for this month, and we've got a theoretical date for the kick-off - namely November 17.

That's the word from the rumor mill, anyway, though we should add a whole lot of salt to this, as the source is MyDrivers (specifically Kuai Technology). Note that MyDrivers is not top of our list of favored sources, it has to be said (but it has made correct predictions in the past).

We're told that 'many' graphics card makers over in China have confirmed to the tech site that the RTX 4090 will be banned in the country on November 17.

NVIDIA's family of AI GPUs are big business... big, big business... so much so that they're sold out into 2024, and prices are skyrocketing because companies small and large want the best AI GPUs for their workloads.

But now, the "Google" of South Korea, Naver, has switched over from using GPUs to CPUs for some of its AI workloads. Naver was already a customer of NVIDIA, but with issues in the supply chain and prices continuing to rise, the company started using Intel CPUs for the AI platform, looking after fraud detection. This is the first time we've seen a company switch from using AI GPUs from NVIDIA to any CPU for artificial intelligence workloads.

In a new report from The Korea Economy Daily, Naver has replaced NVIDIA GPUs with Intel CPUs for the AI workloads used in its Naver Place map platform. Until now, Naver used NVIDIA AI GPUs to run an AI platform that separated fake listings from real ones in response to user queries. The higher prices and long wait times forced Naver to switch to Intel CPUs, according to sources of The Korean Economy Daily.

NVIDIA's next-gen B100 "Blackwell" GPU has entered the supply chain certification stage, one of the important steps of its journey into becoming the truly next-gen AI GPU that will succeed the H100 "Hopper" AI GPU.

We've had news from Chinese and Taiwanese outlets reporting that NVIDIA's new B100 Blackwell AI GPU entering the Supply Chain Certification stage, according to sources of both UDN and CTEE. According to the new reports, NVIDIA is selecting its supply chain partners, with the primary partner being Wistron for production, who will provide NVIIDA with supply substrates required for the next-gen B100 Blackwell AI GPU.

The reports suggested that NVIDIA was looking at Wistron and Foxconn, but Wistron was able to secure the early-stage orders through reservations with yields and other factors that were able to impress NVIDIA enough to put pen to paper. We are to expect NVIDIA's next-gen B100 Blackwell AI GPUs in Q2 2024, where they'll use SK hynix's new HBM3e memory that will push the memory bandwidth numbers through a wall, pushing new limits for AI GPUs in 2024 and beyond.

NVIDIA's new Ada Lovelace GPU architecture has powered the GeForce RTX 40 series graphics cards, but the previous-gen Ampere GPUs in the form of the GeForce RTX 30 series graphics solutions are still in plenty of gaming laptops on the market... but now there's a new Laptop GPU made into a desktop GPU... weird.

Now we're seeing the NVIDIA GeForce RTX 3080 Ti Laptop GPU with 16GB of GDDR6 being turned into a desktop RTX 3080 Ti, which has now been put through its paces in 3DMark testing in the form of the GeForce RTX 3080m Ti. Why are we seeing this now? Because something was created out of the cryptocurrency mining insanity, some companies -- mostly in Asian markets -- used mobile GPUs and re-purposed them into desktop graphics cards.

But why? Because the laptop graphics cards didn't have NVIDIA's anti-crypto Lite Hash Rate algorithm, they worked well for cryptocurrency mining operations. They were cheap to buy as well, and NVIDIA didn't put any stops in place to slow any of this down.

We keep hearing every single drip and drop of information on NVIDIA's upcoming GeForce RTX 4080 SUPER graphics card, with some fresh energy today from leaker "kopite7kimi" on Twitter.

The leaker tweeted out that NVIDIA's new GeForce RTX 4080 SUPER graphics card will be using the full AD103-400 GPU with 10240 CUDA cores in total, 320 TMUs, 112 TOPs, and 64MB of L2 cache. NVIDIA is expected to use a 256-bit memory bus, which should be matched with 16GB of GDDR6X memory, which could be clocked faster than the 22.4Gbps on the RTX 4080 to 24Gbps with new Micron GDDR6X chips.

Kopite7kimi tweeted:

We've been hearing about the ASUS GeForce RTX 4060 Ti graphics card with a built-in M.2 SSD slot, and now it's official: the new ASUS DUAL GeForce RTX 4060 Ti SSD graphics card is here.

The new ASUS DUAL GeForce RTX 4060 Ti SSD graphics card has a built-in M.2 SSD slot that will happily push over 6.5GB/sec sitting on the back of your GPU. You can use most PCIe 4.0 M.2 SSDs on the back of this card, enjoying that super high-performance NVMe SSD performance.

The card will support SSDs up to 80mm long, with the SSD transferring any heat generated into the card's own heatsink using a thermal pad, which provides a super-compact and neat-looking style. ASUS includes the M.2 Q-LATCH, which makes it easy to install or remote the M.2 SSD on the back of the ASUS DUAL GeForce RTX 4060 Ti SSD graphics card without needing any specific tools. There's a design that uses an easy-to-use simple locking mechanism to secure the drive, removing the need for any screws for the SSD.

AMD has released a new version of its Adrenalin graphics driver especially for Alan Wake 2.

Adrenalin Edition 23.20.17.05, which is clearly marked 'For Alan Wake 2,' has been rushed out the door by Team Red.

A dedicated driver just for a single game? Yes, that's unusual for sure, but what seems to have happened here is that AMD couldn't quite fit the work to support Alan Wake 2 into the last driver (v23.10.2, released last week).

Those mulling a purchase of an RTX 4090 graphics card would be advised to wait, at least going by what's happening with the GPU's pricing right now, and also what might be in store for the future (as predicted by leaker Moore's Law is Dead - we'll come back to that later).

Wccftech reports that the pricing of RTX 4090 models in the US is climbing now, after having held pretty steadily at around the MSRP of $1,599 for some time. It's now difficult to find one at MSRP, and if we look at major retailers, their 4090 models are a good chunk above that level now.

The tech site points to Newegg, where the cheapest RTX 4090 is $1766 (at the time of writing, it still is), which is over 10% more than the MSRP, a notable rise.

We've been hearing more and more about NVIDIA's upcoming GeForce RTX 4080 SUPER 16GB graphics card, hearing a few different rumors suggesting AD103 GPU, then AD102, 20GB of GDDR6X, then 16GB of GDDR6X... well, now we're hearing about pricing, and it's exciting.

NVIDIA's upgraded GeForce RTX 4080 SUPER 16GB graphics card could be released to really stir things up with its competitor AMD, not only dropping it into battle above the Radeon RX 7900 XTX, but having a price of possibly just $999 to really make an impact. Moore's Law is Dead reports that a $999 to $1199 price could arrive with the RTX 4080 SUPER 16GB and will be pushed to perform 9-15% faster than the RTX 4080.

In terms of upgraded memory, MLID is being told directly that the RTX 4080 SUPER would be using faster 24Gbps GDDR6X memory compared to the 21Gbps GDDR6X memory on the current RTX 40 series GPUs. If NVIDIA does price its RTX 4080 SUPER at $999 to $1199, then we'd most likely see the RTX 4080 and RTX 4070 Ti discontinued and the RTX 4070 dropping to below $550.

NVIDIA's flagship Hopper H100 Tensor Core GPU is one of the best pieces of technology that mankind makes, costing over $35,000. Alphacool has just unveiled a new single-slot water block for the NVIDIA H100 PCIe GPU that has been designed and engineered for Grace Hopper H100.

Alphacool's new ES H100 80GB HBM PCIe GPU cooler features a matte carbon finish with no RGB bling -- this is a data center, professional, mega-expensive card, not a gaming GPU -- which will see many of these installed into a motherboard. This will also be driven by the fact that Alphacool's new H100 GPU cooler is a tiny single-slot instead of a 1.5-slot design.

The company positions the GPU cooler as close as possible to the components that need to be cooled, using heat-conducting pads that were reduced to a thickness of just 1mm. Alphacool wanted the maximum possible reduction in the thickness of the copper block, and the optimization of the water flow inside the cooler to allow all important components, including the GPU, VRMs, and VRAM, to be cooled by water much better and more effectively. It's a win-win.

There have been rumors about a possible SUPER refresh for some of the GeForce RTX 40 Series line-up. It's a name we haven't seen since the GeForce RTX 20 Series, and if you're a little confused as to how the GeForce RTX 4080 SUPER and GeForce RTX 4070 SUPER fit into the product stack - don't worry, we're about to break it down.

The GeForce RTX 2080 Ti was the only GPU from the Turing generation to sport the 'Ti' branding - with NVIDIA introducing more powerful SUPER versions of the GeForce RTX 2060, RTX 2070, and RTX 2080 in the back half of the generation's life cycle. The SUPER branding didn't appear during the Ampere-based GeForce RTX 30 Series (instead, we got a bunch of 'Ti' models). Still, several rumors indicate its return with the latest Ada Lovelace generation.

There's a lot of info out there regarding the new SUPER line-up. However, some recent comments and info from known industry insider and leaker @kopite7kimi clearly show what to expect. And that is similar to what we saw with the GeForce RTX 20 Series and its SUPER lineup.

ASUS has just teased its first entry from its TUF Gaming family of graphics cards that's white, within the newer NVIDIA GeForce RTX 40 series GPU family. We've seen ASUS unleash Radeon RX 7800 and Radeon RX 7700 XT models in white, but this is a first for the RTX 40 series from the company. Check it out:

The company already has a regular black-themed TUF Gaming GeForce RTX 4070 Ti graphics card, so there are no physical changes to the appearance of the White OC Edition. The shroud, fans, and backplate have all been changed to white, obviously, with no modifications done to the I/O bracket or even the PCB -- the PCB is still black -- so it's the same card, pretty much dipped in white paint.

NVIDIA's in-house GeForce RTX 4070 Ti Founders Edition is different to the ASUS TUF Gaming GeForce RTX 4070 Ti OC Edition in a few ways: there's a regular 8-pin PCIe power connector here, but the biggest change is something I loved during my reviews of the ASUS GeForce RTX 40 series GPUs... ASUS includes an additional HDMI 2.1 connector on their cards, which is quite useful in some scenarios.

Recently, the U.S. government updated its list of graphics hardware that is effectively banned from being exported or sold to China without a government license, including many of NVIDIA's most popular AI GPUs like the A100, A800, H100, and H800.

A new filing from NVIDIA outlines that the new ban is "effective immediately," meaning that sales and exports for these GPUs have been shown a bright red light.

Bad news for AI in the region, but good news for PC gamers? As previously reported, the ban was about restricting hardware sales for GPUs and hardware exceeding a certain performance threshold - which, in a testament to its raw power, included the GeForce RTX 4090. The most powerful gaming GPU on the planet was on the cusp of being banned in one of the biggest PC gaming markets in the world.

Intel Game On Drivers for the Intel Arc range of graphics cards (including the new Intel Arc A580 we recently reviewed) have been released at a steady clip - offering support for new games in addition to delivering massive performance improvements to older titles that didn't play nice with the architecture when it debuted last year.

The latest driver release, 'Intel Graphics Driver 31.0.101.4900 for Intel Arc,' focuses squarely on new titles: Ghostrunner 2, Alan Wake 2, Cities: Skylines 2, and WRC. However, two of these games have been making headlines of late due to just how demanding they are on PC hardware, which raises the question of how they'd run on the entry-level and mainstream Intel Arc lineup of GPUs.

Cities: Skylines II is launching in a state on PC that the developer has called out as unfinished, so much so that some players are struggling to maintain higher than 10 FPS with high-end hardware. So when Intel writes, "Up to 34% average FPS uplift at 1080p with High settings," in the release notes - without context or specific figures, it could be misleading.

ASL has just revealed a graphics card the world simply doesn't need: a new GeForce RTX 4060 with a built-in spirit level.

The new ASL GeForce RTX 4060 graphics card isn't the first with a built-in spirit level, as the 51RISC GeForce RTX 4090 had one, as well as Sycom's custom GeForce RTX 4090 and RTX 4080 cards with AIO cooling and a built-in spirit level. This is a much thinner card and a much more basic design.

ASL's new custom GeForce RTX 4060 has a dual-fan cooler, a single 8-pin PCIe power connector, and a silver-themed design. But the spirit level... it's on the end of the card, on the front next to the two fans. Once the ASL GeForce RTX 4060 is installed into a system, it will not even be visible. So the spirit level on the card itself serves zero purpose, zero use... and then once it's installed, you can't even look at it. If you had a vertical mount for the GPU, then sure... but why.

NVIDIA's mainstream GeForce RTX 4060 Ti 16GB has appeared in a new form: a single-slot, blower-style fan for your new SFF gaming PC.

The new GeForce RTX 4060 Ti 16GB single-slot graphics card was teased by MEGASizeGPU, which will be a release from COLORFUL, but we don't know if it's an iGame GPU yet or not. The new single-slot GeForce RTX 4060 Ti 16GB arrives in a super-tight, single-slot design that measures 267mm long and just 20mm thick.

A blower-style fan is used for cooling, with not many air cutouts at all above the I/O ports -- which are standard here on the GeForce RTX 4060 Ti 16GB single-slot GPU: 1 x HDMI 2.1 and 3 x DisplayPort 1.4 connectors. A single 8-pin PCIe power connector is at the rear, making the card look very neat and tidy all around.

NVIDIA's new GeForce RTX 4080 SUPER refresh GPU is currently in the development oven at Team Green, with some new ponderings about its specifications -- and how it can get around the new US sanctions against China -- which have seen the AD102-based GeForce RTX 4090 banned, and the price of the GPU skyrocketing in China.

The upcoming refreshed GeForce RTX 4080 SUPER has been rumored with 20GB of GDDR6X, but a new video from Moore's Law is Dead suggests that we'll see some changes to the refreshed SUPER card that will skirt around the US sanctions in China... and it totally makes sense for NVIDIA to do this.

The Biden administration pushed in new sanctions on China that have outright banned the GeForce RTX 4090 from being sold in the country, with prices jumping within hours. Moore's Law is Dead is reporting from his sources that NVIDIA will not be moving any GeForce RTX 4090 graphics card stock out of Asia and into any other country in any big numbers because shipping containers are so expensive as we head into the holiday season.

NVIDIA's next moves in the GPU space have been under scrutiny lately, with a whole bunch of recent rumors about possible refreshes of the RTX 4070 and 4080 - and now we're hearing speculation on where Team Green could go with pricing.

This comes from Moore's Law is Dead (MLID), who reasons his way through a theory - and it is just that, theorizing, albeit based on commentary from multiple trusted sources - as to what NVIDIA's pricing strategy might be going forward this year into next.

We've already been hearing about the effects of the US ban on the RTX 4090 GPU in China - well, technically the AD102 chip, we should clarify - and MLID reckons this is going to have, shall we say, interesting ramifications for the graphics card world.

Recently, we reported on the ban on the GeForce RTX 4090 GPU in China, which shocked several people because it's the first time the US-imposed export ban on technology to the country included a piece of hardware designed for PC gaming. This has led to RTX 4090 models being sold for double the price in China, ahead of the restrictions kicking in next month.

According to sources, it's all about restricting and slowing China regarding AI. With the new restrictions looking to limit the sale of NVIDIA's A800 and H800 workstation cards in the region, the ban, which doesn't restrict specific models but hardware with certain transistor numbers, now includes the GeForce RTX 4090.

No GeForce RTX 4090s for China might not sound like a big deal, but with the AI boom and new reports from outlets like TechSpot, it seems that when it comes to high-end NVIDIA GeForce RTX GPUs, we're on the cusp of another situation where prices are about to go through the roof.