Graphics Cards - Page 57

AMD's next-gen Radeon RX 8000 series "RDNA 4" mobile GPU details have leaked: 4 SKUs, 16GB, 12GB, 8GB variants with up to 175W TGPs coming to gaming laptops in 2025.

In a new post by leaker Golden Pig Upgrade, we're being told that the mobile RDNA 4 GPU offerings will arrive in two Navi 4X GPU camps: Navi 48 and Navi 44. We can expect two Navi 48-based SKUs which should materialize into a Navi 48 GPU with 16GB GDDR6 memory on a 256-bit memory bus, and a Navi 48 SKU with 12GB of GDDR6 on a 192-bit memory bus.

Under that, we're to expect two Navi 44-based mobile RDNA 4 GPUs, both with 8GB of GDDR6 on a 128-bit memory bus. The bigger difference between the Navi 44 mobile GPUs is that the "P25M-P6" variant will have a 75-150W TGP, while the "P25M-P4" variant will have a 50-130W TGP range. These new mobile RDNA 4 GPUs will launch in 2025, 2025 inside of new gaming laptops.

NVIDIA hasn't even made its next-gen GeForce RTX 50 series "Blackwell" GPUs official yet, but the company is full-steam ahead trademarking future-gen GeForce RTX 60 series GPUs including new flagship cards like the GeForce RTX 6090 SUPER, RTX 6090 Ti, and RTX 6090 graphics cards.

In a new submittion of new RTX series GPUs to the Eurasian Economic Commission (ECC) regulatory body, NVIDIA is preparing itself for a wave of cards including the GeForce RTX 6090 SUPER, RTX 6090 Ti, and RTX 6090 as well as the RTX 5090 SUPER, RTX 5090 Ti, and RTX 5090. This is at the top end of each series, but there are filings for the RTX 6080 SUPER, RTX 6080 Ti, and so on.

We shouldn't expect all of these GPUs to come out, but NVIDIA has to prepare ahead of time as AMD is cooking behind the scenes with its next-gen RDNA 5 GPU architecture (the one after RDNA 4 which launches in 2025). We should expect NVIDIA to dub its future-gen GeForce RTX 60 series "Rubin" after the Rubin R100 AI GPU (with next-gen HBM4 memory) launches in 2025.

We're on the cusp of a new GPU generation for PC gaming, and all signs are pointing to CES 2025 in Las Vegas in January as the place where we will see everything for the first time. NVIDIA is prepping to debut its GeForce RTX 50 Series, powered by the new Blackwell architecture, led by the GeForce RTX 5090, RTX 5080, and RTX 5070.

AMD is set to announce its new line-up of RDNA 4-powered cards with what we assume will be the mid-range Radeon RX 8800 XT and Radeon RX 8700 XT combo. CES 2025 is also where many assumed Intel would debut its second generation of desktop Arc graphics cards with new mainstream Battlemage 'Xe2' products.

It's no secret that Intel's first Arc GPUs had a rocky launch regarding driver stability and performance and have failed to capture any meaningful market share from NVIDIA or AMD. A new leak suggests that Intel is prepping for a Battlemage debut in December, ahead of CES 2025.

MSI is getting in on the action with discounted NVIDIA RTX 4000 GPUs - as Lovelace prepares to be succeeded by new Blackwell graphics cards - and there's a tasty offer on an RTX 4070 Ti Super.

MSI's RTX 4070 Ti Super Expert model is now on sale direct from the company at $729.99, which represents a chunky discount. This GPU normally retails at $849.99 from the MSI store, so you're getting $120 off.

As you might expect, there's limited stock, but there is still inventory available at the time of writing.

NVIDIA's next-generation GeForce RTX 50 series "Blackwell" gaming GPUs aren't far away, with a reputable leaker saying "we will meet GeForce or Blackwell soon" on X just hours ago. Check it out:

The post is coming from leaker "kopite7kimi" with a simple "We will meet GeForce of Blackwell soon" which goes against previous rumors that NVIDIA would be introducing its new GeForce RTX 50 series GPUs at CES 2025 in January. We had heard not too long ago that NVIDIA was reportedly preparing the GeForce RTX 5090, RTX 5080, and RTX 5070 reveals at CES 2025 with releases staggered through January, February, and March 2025.

NVIDIA has hosted pretty large events for each new generation of GeForce graphics cards, so a big reveal at CES 2025 while a million-and-one other things are being announced, means that Team Green won't be the star of the show.

NVIDIA currently controls over 85% of the global GPU market, and its controlling share of the gaming GPU market is hovering around 90%. That is hundreds of millions of NVIDIA GPUs globally, and now the company has confirmed it has identified at least eight "high-risk" security vulnerabilities within its GeForce GPU drivers.

While GeForce typically means gaming GPUs specifically, that isn't the case here as professional and workstation GPUs have also been listed as victims of the vulnerabilities. NVIDIA has taken to its Security Bulletin to outline each of the vulnerabilities, their risk, level on the severity scale, and the fix. According to the website, the flaws were within NVIDIA's GPU Display Driver and the NVIDIA vGPU Software.

NVIDIA writes each of the vulnerabilities could enable a hacker to execute code, escalate administrative privileges, information harvesting, data tampering, and a complete denial of service. Notably, each of the vulnerabilities were graded on a severity scale from 0 to 10. The lowest score out of the eight that was ranked was 7.1, and the highest was 8.2, indicating these vulnerabilities weren't small issues within NVIDIA's GPU drivers as internally they were graded quite high.

The latest Steam Hardware & Software Survey Results for October 2024 are in, and when it comes to the GPUs PC gamers are using, it's a sea of green. According to Valve, NVIDIA's GeForce graphics cards have a 77.37% share of the PC gaming market, and the top 34 discrete GPUs (covering mobile and desktop graphics cards) are all GeForce products.

The most popular desktop Radeon option, the Radeon RX 6600, appears in the number 35 spot ahead of the Radeon RX 580. Both cards are at least a few generations old, and even though that covers most of the list - the most popular gaming GPU is still the GeForce RTX 3060 - it paints an increasingly bleak picture of the current state of Radeon.

On the plus side, AMD finally has two RDNA 3 options in the list - the AMD Radeon RX 7900 XTX and AMD Radeon RX 7700 XT - which both appear toward the bottom. On the flip side, NVIDIA's latest GeForce RTX 40 Series has five entries in the Top 20, showcasing that the Ada Lovelace generation has been another success for NVIDIA.

Intel's troubles have been widely publicized and reported throughout the year, from CPU launch issues to behind-the-scenes development and foundry production woes. It's worth keeping this in mind because, with layoffs and cost-cutting measures in place, Intel's current and future decisions will most likely be influenced by its current state.

Okay, with that out of the way, let's look at the future of Intel Arc based on comments made by Intel CEO Pat Gelsinger during a recent Q3 2024 Earnings Call. Essentially, with next-gen Arc Xe2 architecture graphics becoming a part of the company's Lunar Lake mobile line-up - the future of Intel's entry into the discrete graphics market will focus on integrated graphics.

"We've taken steps to simplify the product line, have fewer SKUs to cover the marketplace, and we're focused on the efficiencies associated with that," Pat Gelsinger said. Simplifying the road map means"large integrated graphics capabilities" will become the new norm for Intel, with "less need for discrete graphics in the market going forward."

Minisforum is a company known for creating powerful Mini PCs for various users, from gamers to everyday users. Its latest is a stylish external GPU docking station that features the mobile AMD Radeon RX 7600M XT graphics card with 8GB of VRAM.

The Minisforum MGA1 is a portable OCuLink eGPU with a stylish design resembling a more minimal version of the similarly specced AYANEO AG01 that launched last week. The Minisforum MGA1 is a little cheaper, at $559 USD, compared to the AYANEO AG01's $599 price point.

The MGA1 is designed to be paired with Mini PCs and gaming handhelds with an OCuLink port and includes 1 x HDM1 2.1 and 2 x DisplayPort 2.0 video output ports for up to 8K at 60 Hz. It also has three USB 3.2 10Gbps ports and a single USB 3.1 Type-C port with 65W power delivery.

NVIDIA's DLSS 3.5 Ray Reconstruction, which debuted in 2023, has been a game changer for PC games with multiple ray-tracing effects - particularly those with ray-traced reflections. It's an AI-powered denoiser, a part of the rendering process that cleans up an image.

NVIDIA's Ray Reconstruction AI model was trained on over 6X the data used for DLSS Super Resolution and Frame Generation. It's a game-changer that 'fills in the gaps' in a way that dramatically improves the image quality of ray-traced effects. Reflections become sharper and more detailed, and the lighting looks more realistic and cinematic.

The difference is so big that games with ray tracing look better on GeForce RTX rigs - plain and simple. The good news is that AMD is working on a similar-sounding AI denoiser that could be part of RDNA 4's new "AI capabilities." Also, its FSR super-sampling or upscaling is going AI.

Yesterday, we reported on AMD's Q3 2024 financial results, which showed revenue and growth once again driven by its Data Center segment. Regarding AMD's Gaming business, which covers Radeon products for PC and console production for Sony and Microsoft, revenue was down 69% - with extremely thin profit margins.

As part of the earnings call, AMD CEO Lisa Su addressed the slump in the company's gaming business by teasing its next-generation RDNA 4 products. She confirms that RDNA 4 GPUs are on track to launch "in early 2025" and has teased what to expect.

In addition to a "strong increase in gaming performance," Lisa Su has confirmed two RDNA 4 features for next-gen Radeon RX 8000 Series GPUs that should bring its capabilities closer to the NVIDIA and GeForce RTX juggernaut.

AMD supposedly has a whole heap of revelations set for CES 2025, and that includes new RDNA 4 graphics cards, and Strix Halo APUs, among many other goodies.

On X, Everest flagged up a fellow leaker from Chiphell, Zhangzhonghao, who aired a long list of the products AMD wants to wow us with at the big Consumer Electronics Show in January.

In the laptop arena, that includes top-end Strix Point Halo APUs, plus 'Krackan Point' (Kraken Point), as well as next-gen Fire Range chips (to succeed Dragon Range).

On Friday, October 25, 2024, NVIDIA briefly overtook Apple to become the world's most valuable company with an eye-watering market cap of $3.53 trillion. This put NVIDIA's valuation higher than Apple's $3.52 for a moment, closing the day at $3.47 trillion in second place ahead of Microsoft.

This market marks the second time NVIDIA has taken the crown from its tech giant rivals Apple and Microsoft, reaching this height in June 2024. Even though it feels like we're only starting to get used to talking about NVIDIA in the same light as companies like Apple, Microsoft, Google, and Meta, with everyone laser-focused on generative AI, this remarkable result makes sense.

NVIDIA, the clear leader in AI hardware software and chip design, has hardware and technology in almost every data center and system currently working with AI models, from significant endeavors from companies like OpenAI to workstations sitting on a desk running desktop RTX hardware in the office of a small startup somewhere.

AMD has just fired off an email blast that claiming its Radeon graphics cards offer up to 38% more performance than the competition (NVIDIA) in Call of Duty: Black Ops 6.

The Red Team states in its email that AMD Radeon RX 7000 series graphics cards can offer gamers up to 38% more performance than the competition, which at this point is just NVIDIA, considering Intel owns less than one percent of the graphics card market. It could even be argued that AMD isn't much competition, considering NVIDIA's dominance at 88% market share of all GPU shipments.

Regardless of those figures, AMD states in its e-blast that even at the "most extreme settings," it will beat the competition in terms of performance in Call of Duty: Black Ops 6. Moreover, AMD highlighted its alleged wins and how each Radeon card would offer the claimed performance increase over the competition. According to the Red Team, the Radeon RX 79000 XTX and the RX 7900 XT offer up to 22 and 21% faster performance than "competing graphics solutions at native 4K with Extreme settings enabled".

NVIDIA's upcoming GeForce RTX 50 series Laptop GPUs have been leaked, revealed at the PCI ID Repository, giving us a clearer look at the mobile Blackwell GPUs deploying in 2025.

The NVIDIA GeForce RTX 50 series "Blackwell" GPU lineup features both desktop and laptop variants, with the main focus being on the RTX 50 series desktop GPUs, but now we're learning about each of the SKUs of RTX 50 series Laptop GPUs, and there's quite a lot.

We've been hearing through previous rumors that NVIDIA's flagship GeForce RTX 5090 Laptop GPU would feature 24GB of GDDR7 memory on a 256-bit memory bus, but we could still expect to see the RTX 5090 Laptop GPU launch with 16GB of GDDR7 memory instead. The RTX 5080 Laptop GPU should also have 16GB of GDDR7, while the rest will have 12GB or 8GB of GDDR7 memory.

The new AYANEO AG01: Starship is the company's first eGPU or external graphics solution and is now available for purchase. Shipping is slated for the end of November. For the $599 USD price, you get a portable AMD Radeon RX 7600M XT GPU with 8GB of DDR6 memory that can connect via OCuLink, USB4, or Thunderbolt.

Additional features and functionality of the device include additional USB-A and Ethernet ports, a slot for an M.2 2280 SSD, SD Card expansion, and multiple HDMI and DisplayPort outputs. Designed for gaming handhelds, laptops, and Mini PCs, the AYANEO AG01: Starship is an all-in-one gaming dock for various devices. The SSD slot is an excellent touch as the AYANEO AG01: Starship can store all of the games it's used to play on the device.

Setting all of that aside, the physical design and look of the AYANEO AG01: Starship is impressive, like something out of a sci-fi film from the 1980s. The big round button is not only the "engine" start for the unit, but it also includes a dial to boost and overclock the GPU by upping the power from 100W to 120W.

NVIDIA's latest Game Ready Driver comes with the usual features, including support for some major new PC games - Call of Duty: Black Ops 6 is one of them - but it packs a sting in the tail for those with Corsair peripherals that sport RGB lighting.

Mainly because NVIDIA's driver v566.03 has a conflict with Corsair's iCUE tool, used for controlling its RGB lighting (and other functionality besides).

Tom's Hardware pointed out the problem (acknowledged by Corsair as you can see above, in its post on X) whereby CPU usage is spiking not just with iCUE, but also Bluestacks, the popular Android emulator for PC.

Most gamers only really care about how many frames per second their graphics card can push at what graphical settings, but what GPUs and the impressive number-crunching technology insider is capable of much more than just hitting maximum frame rates in games.

Perhaps a more unusual example of the power of a GPU comes from a former NVIDIA engineer who has decided to use a NVIDIA A100 GPU to discover what is now considered to be the largest prime number. According to a new report from The Register, the new prime number has been named "M136279841," and it's equal to the monstrous number that is 2 to the power of 136,279,841 - 1. For those interested, the Great Internet Mesrsenne Prime (GIMPS) Search has a ZIP file containing a text file that has every single one of the 41,024,320 digits.

The download size of the file, or the largest prime number ever discovered, is just shy of 20MB at 18.3MB, which is substantially large for a text file. Former NVIDIA engineer Luke Durant has spotted what is called the 52nd Mersenne prime, which is a prime number that equals 2n - 1. According to the report, the last time a Mersenne prime was discovered was by Patrick Laroche more than five years ago, who found it by running the free GIMPS software on a four-core Intel Core i5-4590T processor over 12 days.

The upcoming Blackwell "GB203" GPU die has been leaked, the GPU that should power NVIDIA's next-gen GeForce RTX 5080 desktop graphics card, and the RTX 5090 Laptop GPU and RTX 5080 Laptop GPU. Check it out, in all its TSMC 4nm glory:

In a new video, leaker Moore's Law is Dead has posted a purported GB203 die shot which will power the second-fastest Blackwell desktop gaming GPU -- the GeForce RTX 5080 -- and what will be the two best discrete GPUs inside of laptops in 2025: the GeForce RTX 5090 Laptop GPU and the GeForce RTX 5080 Laptop GPU. We'll also have next-generation GDDR7 memory across the board.

This is just a leak, so the GB203 GPU die that we're looking at might not be 100% making its way into the RTX 5080, or the RTX 5090 Laptop + RTX 5080 Laptop GPUs, it could very well make it into a laptop-focused Blackwell-based workstation GPU. We only have a couple of months to wait until we get some confirmation, probably weeks at this rate.

NVIDIA's latest GeForce Game Ready Driver, Verison 566.03, is a big one, adding day-one support for a massive list of new PC games: Alan Wake 2: The Lake House, Call of Duty: Black Ops 6, Dragon Age: The Veilguard, Horizon Zero Dawn Remastered, No More Room In Hell 2, Red Dead Redemption, and The Axis Unseen.

In addition, it includes support for 32 new G-SYNC-compatible displays, covering a range of new Samsung Odyssey monitors and several new OLED TVs from Panasonic and Philips. Plus, new one-click Optimal Settings for DRAGON BALL: Sparking! ZERO, NBA 2K5, and Silent Hill 2 in the NVIDIA App or GeForce Experience.

As the holiday season approaches and several big game releases creep up, now's the time to keep your GPU drivers up to date to ensure the best performance. Especially for a game like Dragon Age: The Veilguard, which launches on October 31 with DLSS 3, Reflex, and ray tracing support.