Graphics Cards - Page 62

AMD's next-gen RDNA 4-based Navi 48 GPU leaks suggest we can expect gaming performance of RTX 4080 levels, with far better ray tracing performance out of RDNA 4, too.

In his latest video, leaker Moore's Law is Dead notes that the rasterization performance is "expected to at least trade blows with the RTX 4080" while ray tracing performance is expected to "trade blows with the RTX 4070 Ti SUPER, and in some AMD optimized games, there's suggestion that it could get close to the RTX 4080!"

The leaker continues, explaining that Navi 48 will have "high volume" in Q4 2024, and that the company "hopes it will be a hit with system integrators". The new Radeon RX 8800 XT (or whatever it is called) will "have a reference cooler, but not be sold on AMD.com".

NVIDIA quietly announced the release of a new GeForce RTX 4070 variant that features the slower and cheaper GDDR6 memory. NVIDIA states the RTX 4070 GDDR6 variant will offer "similar" performance to its GDDR6X cousin.

Unfortunately, NVIDIA hasn't mandated any clear packaging guidelines for AIBs to inform buyers of the new card has the slower memory, as name "GeForce RTX 4070" features the exact same name as the GDDR6X variant. The lack of packaging change makes it extremely difficult for buyers that aren't familiar with technological terminologies to make informed purchasing decisions, especially considering NVIDIA's decision to price the GDDR6 variant at the same price point as the GDDR6X variant.

NVIDIA stating the new card will have "similar" performance means it won't be identical performance to the GDDR6X variant, which means it will likely fall within less than 5% on average of the GDDR6X card. Moreover, the GDDR6 variant doesn't have any special naming such as "GDDR6" on the box, or feature any monkier title such as "SUPER" or "Ti" to distinguish it from the original RTX 4070.

NVIDIA's next-generation GeForce RTX 5060 Laptop GPU is rumored to have gaming performance at levels of the RTX 4070, while using less power.

In some new details from leaker Golden Pig Upgrade, who attended an event held recently by Chinese laptop and PC manufacturer, Shenzhen Hasee Compute Co. where during the event, the chairman of the company, Wu Haijun, talked about next-gen PCs.

During that chat, he talked about some details about NVIDIA's next-generation GeForce RTX 50 series "Blackwell" GPUs. The Hasee chairman said that NVIDIA's next-gen GeForce RTX 50 series Laptop GPUs will use GDDR7 memory, and power consumption of the 60-level cards will be "greatly reduced" from 140W to 115W.

AMD's next-generation RDNA 4 GPU architecture has been teased again, with an RDNA 4-based Radeon RX 8000 series graphics card turning up on Geekbench.

The new graphics card in question is an RDNA 4-based GPU with 56 CUs (Compute Units) and 16GB of VRAM (which should be GDDR6 still, but faster 18Gbps GDDR6 memory chips). The GPU has a "GFX1201" device ID, which means we're looking at the Navi 48 SKU.

AMD's new Navi 48 "GFX1201" GPU is listed as 28 compute units, which results in 56 compute untis as the new RDNA 4 GPU architecture does things the same as RDNA 3: each GPU has a shader engine that features dual compute units, so we should see a doubling of the listed compute units (28 x 2 = 56).

The upcoming Call of Duty: Black Ops 6 Beta timings have yet to be confirmed, but it's expected sometime soon. The full game is out on October 25 for PC, PlayStation, and Xbox - so we can expect that multiplayer first look soon. It could be here in a matter of days as the latest Intel Graphics Driver 32.0.101.5972 (WHQL Certified) has added support for the Call of Duty: Black Ops 6 Beta for Intel Arc graphics owners across mobile and PC.

Intel's latest driver also adds day-one support and optimizations for SMITE 2, Star Wars Outlaws, and Visions of Mana. The company added support for Black Myth: Wukong to its most recent non-WHQL driver release. Weighing in at 884.4 MB, Intel Graphics Driver 32.0.101.5972 is available now for Windows 10 and Windows 11 systems with discrete Intel Arc graphics cards and Intel Core Ultra devices with built-in Arc graphics.

Support for these new PC games looks to be the extent of the latest driver release, with several 'Known Issues' remaining unresolved. Here are the full release notes.

Intel has just published the first public driver for Xe2 graphics, with the release of this driver important because its next-gen Core Ultra 200V series "Lunar Lake" processors are launching next month.

We've heard through leaks that the Xe2 codename "Battlemage" GPUs would be called Arc 140V and Arc 130V, and it looks like the leaks were right, as Intel's own information lists support with the new Arc Graphics 32.0.101.5972 drivers that "Arc 140V" and "Arc 130V" GPUs are supported.

According to the last leaks, Intel's new Core Ultra 9 and Core Ultra 7 "Lunar Lake" processors will get Arc 140V with 8 Xe2-Cores, while the Core Ultra 5 SKUs will get Arc 130V with 7 Xe2-Cores. The highest-end Core Ultra 9 288V will have its Arc 140V GPU boosting up to 2.05GHz, while the lowest-end Lunar Lake chip will have its Arc 130V integrated GPU boosting at up to 1.85GHz.

ASUS has partnered with AMD and Focus Home Interactive to create a custom GPU to celebrate the PC launch of Warhammer 40:000: Space Marine 2. It's a very tasty-looking AMD Radeon RX 7900 XTX. The 7900 XTX is AMD's current flagship gaming GPU and a powerful card capable of delivering triple-digit 4K performance in several titles.

This limited-edition GPU, which only has 20 units, is a custom version of the company's TUF Gaming design. With Space Marine blue and other little Warhammer 40,000 touches, it's a fantastic-looking card. With ASUS confirming at Gamescom over the weekend that it won't be available to purchase, all 20 units will be given away via this Warhammer promotion website.

Warhammer 40:000: Space Marine 2 is out on September 9, 2024. It will support AMD's updated FSR 3.1 for upscaling and frame generation on PC. FSR 3.1 brings improvements to both FSR's upscaling and frame generation image quality, with support for all modern graphics cards

The underlying technology of all of these AI-powered tools that are continuous popping up in various sectors of the technology industry are Large Language Models (LLMs), which are powered by number-crunching graphics cards, or GPUs.

The explosion of AI has been fuelled by the possibility of graphics cards now being powerful enough to crunch large swaths of data into a model that can then be queried by users. However, powerful models such as ChatGPT, require thousands of GPUs to continuously fulfil all the requests of the millions of users enjoying the service.

But what if you just wanted to support a few thousand users at a time? Perhaps you are a business that could need an AI chatbot to assist with customer service via your website? And ultimately, do you need thousands of GPUs to achieve this? Apparently not, or at least according to benchmarks from Backprop, an Estonian GPU cloud start-up who managed to get a modest LLM such as Llama 3.1 8B to work on a single NVIDIA RTX 3090 GPU, which was released in late 2020.

World of Warcraft is celebrating its 20th anniversary this year. Blizzard's celebrated massively multiplayer RPG is still going strong, with the game set to get its tenth expansion, World of Warcraft: The War Within, next week. It kicks off the new Worldsoul Saga trilogy, with the story being led by veteran Warcraft and Blizzard writer and designer Chris Metzen.

To help celebrate, MSI has partnered with Blizzard to create a custom GeForce RTX 4070 SUPER 12G GAMING SLIM World of Warcraft Edition graphics card with removable 'puzzle-style' magnetic backplates that feature artwork from 20 years of WoW expansions. If you're a World of Warcraft fan, this might be the ultimate GPU.

The GeForce RTX 4070 SUPER is the current mid-range GPU champ, delivering excellent 1440p gaming performance with access to NVIDIA's cutting-edge RTX technologies like DLSS. MSI's GAMING SLIM design is also one of our favorites, and it looks great in gold and blue.

According to a new report from Moore's Law is Dead, citing sources directly within AMD, the company could be releasing its flagship RDNA 4 gaming GPU before the end of the year. Tentatively called the Radeon RX 8800 XT, there's a chance it could be revealed or teased this weekend at Gamescom.

Not only that, but performance-wise, it could be an absolute game changer that shakes up the GeForce RTX-dominated PC gaming market - dramatically improved raytracing performance thanks to an increase in RT hardware per Compute Unit, plus 'rasterization' performance (non-RT or upscaled) that will 'trade blows' with the GeForce RTX 4080.

Of course, we are chalking this up as rumor for now, so take this all with a grain of salt. The flagship Navi 48 GPU (AMD won't be competing on the high-end with its upcoming RDNA 4 generation of graphics cards) is set to feature 64 CUs or Compute Units with Boost Clock frequencies that could hit up to 3.2 GHz.

Intel says that it is ready to 'take our second crack' at the discrete GPU market with Battlemage, its new Xe2 GPU architecture that will debut inside of Lunar Lake-based laptops later this year.

In the latest episode of Intel's in-house "Talking Tech" series, Tom Peterson (or TAP as we call him in the industry) talked about what we can expect from Intel's next-gen Xe2 GPU architecture. The new Xe2 "Battlemage" GPU will feature a 50% performance upgrade, a bunch of architectural change, and the lessons learned from Xe "Alchemist".

TAP talked about the basis of Xe2 for Intel graphics is Lunar Lake, which is an mobile processor coming later this year, but the Xe2 Tile will be used for Battlemage: the company's next-generation Arc series discrete graphics cards.

At Gamescom this year, NVIDIA showcased a range of new and upcoming PC games set to go full RTX On with ray-tracing, DLSS, and Reflex technologies. These include the 'game of the moment,' Black Myth: Wukong (check out our video covering the first 20 minutes of the game here), and Indiana Jones and the Great Circle from Machine Games and Xbox.

The latter is out in December and, on PC, will feature 'Full Ray Tracing' and DLSS 3.5 Ray Reconstruction running on the latest id Tech engine. To celebrate both titles, NVIDIA is giving away a pair of custom GeForce RTX 4080 SUPER Founders Edition graphics cards across its social channels - a Black Myth: Wukong and an Indiana Jones and the Great Circle edition.

These are both one of a kind, with the Black Myth: Wukong variant sporting custom wrap art that looks hand-painted, adding a dose of silver and brushed metal with the game's protagonist seemingly hand-drawn onto the faceplate. It looks fantastic, and one lucky PC gamer will get it.

NVIDIA has announced its latest GeForce RTX gaming bundle. For eligible purchases of a desktop GeForce RTX 40 Series GPU or GeForce RTX laptop, get a free copy of Ubisoft's Star Wars Outlaws with a bonus Forest Commando Character Pack. The pack includes a complete cosmetic outfit for the game's protagonist, reminiscent of Leia's outfit during the Battle for Endor in Star Wars Episode VI: Return of the Jedi.

On the GPU side, qualifying cards include the GeForce RTX 4070, RTX 4070 SUPER, RTX 4070 Ti, RTX 4070 Ti SUPER, RTX 4080, RTX 4080 SUPER, and the flagship GeForce RTX 4090. Unfortunately, GeForce RTX 4060 and RTX 4060 Ti cards are not eligible for the promotion, which is a shame. Getting a free AAA game with the $299 USD GeForce RTX 4060 would have been a sweet deal.

It's a similar story on the laptop front, with qualifying systems having to include at least a GeForce RTX 4070 Laptop GPU.

Although it's described as delivering high frame-rate gaming performance in an SFF build, the new ZOTAC GAMING GeForce RTX 4070 Blower GPU's old-school design probably works just as well in a workstation rig running generative AI tasks. The two-slot design sports a traditional blower fan cooling design with a single intake and hot air exhausted through the vents on the rear.

Zotac isn't the first company to release a GeForce RTX 4070 GPU in a compact two-slot form factor with blower fans; similar models are available from companies like ASUS, GIGABYTE, and others.

The GeForce RTX 4070 is a perfect candidate for this type of design, especially in the AI PC era, as the single 8-pin power connector and 200W TDP with 12GB of GDDR6X memory make it a great cost-effective solution for generative AI workloads. And one that will scale easily by adding more GPUs into the mix.

Zotac Gaming is launching a custom limited edition version of its GeForce RTX 4070 SUPER Trinity graphics card to celebrate the launch of Black Myth: Wukong, which is out this week. Per Videocardz.com, which broke the story and provided the images you see here, this custom GPU will be limited to the Chinese market - a shame.

Zotact's Trinity design is an OC version of the GeForce RTX 4070 SUPER with a faster out-of-the-box boost clock of 2505 MHz compared to the default 2475 MHz. The GeForce RTX 4070 SUPER debuted earlier this year and includes 12GB of GDDR6X memory and a larger 7,168 CUDA Core count compared to the non-SUPER variant.

Our reviews show that it's enough to deliver a notable and impressive bump to 1440p gaming performance. With DLSS 3 and the latest RT hardware, this is the sort of GPU you'll need to play Black Myth: Wukon with its cutting-edge ray tracing effects and stunning character and world detail.

OneXPlayer has just unveiled some new details about its upcoming OneXGPU 2: a new external GPU solution powered by AMD's new Radeon RX 7800M GPU, that's capable of playing AAA games at 60FPS+.

The new OneXGPU 2 is an external GPU that packs the latest USB4 and OCulink connectivity standards, and now we have some more details on its gaming performance. Inside, it's powered by the new Radeon RX 7800M GPU which is based on the Navi 23 die with 60 RDNA 3-based Compute Units for a total of 3840 stream processors, making it equivalent to the desktop RX 7800 XT.

AMD's new Radeon RX 7800M is tweaked on the memory side of things, with 12GB of GDDR6 memory on a 192-bit memory bus, while the desktop RX 7800 XT has 16GB of GDDR6 on a wider 256-bit memory bus. The new RX 7800M GPU has support for up to 3 displays and a maximum TDP of 180W, while GPU clocks will be lower than the desktop variant, obviously.

Arm is working on its own consumer gaming GPU family of graphics cards, which would compete with the likes of AMD, NVIDIA, and Intel.

In a new post from Israeli website Globes, we're learning that Arm has a large team of up to 100 or so GPU and software engineers, at its development center in Ra'anana, Israel. Arm isn't new to the world of GPUs, as it has its in-house Mali and Immortalis series, which are offered to chipmakers as IP to license for their own chips, just like it does with its Arm-based CPUs.

Globes reported: "UK chip giant ARM is developing a graphics processor unit (GPU) in Israel that will compete with NVIDIA and Intel, sources familiar with the matter have told"Globes."ARM is estimated to be employing about 100 chip and software development engineers, in its global graphics processing group at its development center in Ra'anana. At this stage, ARM is reportedly engaged in graphic processing for the video game market, and at the same time - as was the case with NVIDIA - the technology could also be used for AI processing, if and when ARM decides to fully enter the field".

NVIDIA has just introduced its new RTX 2000E Ada workstation graphics card, featuring a single-slot design and 16GB of GDDR6 memory with ECC support.

The company unveiled its new RTX 2000 Ada entry-level workstation GPUs earlier this year, but the new RTX 2000E Ada is something different: the "E" means it has ECC memory. The new NVIDIA RTX 2000E Ada is a low-profile, single-slot workstation GPU with ECC memory that will detect and fix single-bit errors in real-time.

ECC memory is a big deal in the professional world, where processing data accuracy simply can't be compromised, and that's exactly with the new entry-level RTX 2000E Ada workstation GPU is here for. Inside, it features 22 RT cores, 88 Tensor cores, and 2816 CUDA cores.

Last week, we got word that AMD wasn't entirely done with the RDNA 3-powered Radeon 7000 Series of graphics cards, with the potential release of entry-level Radeon RX 7400 and Radeon RX 7300 GPUs set to compete with the likes of NVIDIA's GeForce RTX 3050. The news came from a reliable source. However, a follow-up report now states that AMD and its partners are all-in on RDNA 4.

This means Acer, ASUS, ASRock, Biostar, GIGABYTE, MSI, Powercolor, Sapphire, and XFX are developing and testing new Radeon RX 8000 Series GPUs. As spotted by Videocardz.com, Taiwanese outlet Benchlife notes that none of AMD's Radeon partners are working on entry-level RDNA 3 GPUs.

"Everyone is currently focused on the next generation of Radeon graphics cards based on the RDNA 4 GPU architecture," the Benchlife report reads.

Beelink has just unveiled its latest external GPU dock, which features a 600W built-in PSU and enough juice to support the flagship NVIDIA GeForce RTX 4090 graphics card.

The company isn't the first -- nor will it be the last -- with an external GPU dock, but Beelink is doing something a little different here. Beelink's new external GPU dock features a direct PCIe 4.0 x8 interface, with a physically full-width x16 connector (but only uses half of the lanes) that directly link the eGPU to the Mini-PC motherboard.

Beelink's use of the PCIe 4.0 x8 interface means there's up to 128Gbps (16GB/sec) of maximum theoretical bandwidth, which is much higher than Thunderbolt 4 at 40Gbps (5GB/sec) and even OCulink with 64Gbps (8GB/sec).