Graphics Cards - Page 76

AMD has updated its FidelityFX Super Resolution 3 (FSR 3) technology on GPUOpen to include the full source code for DirectX 12 and Unreal Engine 5 implementation for developers. FSR 3, which includes Frame Generation support for a wide range of AMD Radeon RX and even NVIDIA GeForce RTX GPUs, is the company's answer to the wildly successful and popular DLSS 3 that debuted with the GeForce RTX 40 Series.

FSR 3's rollout has been slow, with only a handful of titles supporting AMD's version of Frame Generation that, like its Super Resolution technology, doesn't require specialized AI hardware. But, FSR 3's most recent appearance - ion Ubisoft's Avatar: Frontiers of Pandora - showcased the best version of the technology to date.

Not only with improvements to overall image quality but also support for Variable Refresh Rate (VRR) displays - an omission that baffled reviewers and gamers when the tech launched in Square Enix's Forspoken and EA's Immortals of Aveum.

Intel Arc 'Game On' driver updates and releases over the past year have been a string of good news stories for the company's first foray into discrete graphics cards for PC gaming. From a massive boost in DirectX 11 performance that has seen double-digit improvements to smoothness and overall frame-rate performance, there has also been a gigantic 2.7X increase in AI performance.

With the launch of the new Meteor Lake-powered Core Ultra series of processors for laptops overnight - which are available with integrated Intel Arc graphics, many have wondered if Intel's steady stream of driver releases for day-one game support and performance increases for older titles will be available for those that pick up a new Core Ultra laptop.

The answer is yes, and this can be seen with the latest Intel Graphics Driver release for Intel Arc GPUs (31.0.101.5081 WHQL and 31.0.101.5122 WHQL), which includes support for the Intel Core Ultra processor family.

GeForce GTX is a brand that has been around for decades, signaling high-end performance for PC gaming that pushed technology and the visual fidelity of games to new heights. Although the company used 'GTX' a few times throughout its history, the first official GeForce GTX GPU to launch was the GeForce GTX 260 in 2008. Since then, the branding has powered mainline NVIDIA gaming hardware until the arrival of AI and hardware-accelerated ray-tracing in 2018.

NVIDIA's desktop and mobile graphics for PC gaming now carry the GeForce RTX branding, and even though it's taken a few years - it's now the go-to GeForce naming for PC gamers. Once again, it pushes technology and the visual fidelity of games to new ray-traced heights.

But, even though GeForce RTX is the new hotness for PC gaming, NVIDIA has still been producing affordable GeForce GTX products, including the GeForce GTX 1660 SUPER, GTX 1650, and low-end GeForce GTX 1630. According to a new report, NVIDIA will stop producing GeForce GTX GPUs sometime in Q1 2024 - marking the end of an era.

NVIDIA is set to have quite the SUPER month in January 2024, with the impending release of its new GeForce RTX 4080 SUPER, RTX 4070 Ti SUPER, and RTX 4070 SUPER cards... but now we've got some release dates for those new refreshed GPUs.

The folks over at ITHome have combined rumors from multiple sources to get some dates penciled in for next month, as NVIDIA will be revealing its new GeForce RTX 40 SUPER series graphics cards at CES 2024 during its "Special Address" event on January 8. The staggered release dates will ensure NVIDIA gets the most eyeballs on its cards instead of dumping them all out on the same day.

The reports suggest that we'll see NVIDIA release the GeForce RTX 4070 SUPER first on January 17, the GeForce RTX 4070 Ti second on January 24, and finally, the GeForce RTX 4080 SUPER on January 31. We don't know anything about GPU clock speeds or exact configurations of the cards, but we do expect the RTX 4070 SUPER to feature 12GB of VRAM and the AD104 GPU with 7168 CUDA cores.

ASAKA has just revealed its new G-Nexus P24 adapter for NVIDIA's current-gen GeForce RTX 40 series GPUs that use 3 x 8-pin PCIe power connectors that support up to 600W of power flowing through them.

The new ASAKA G-Nexus P24 adapter and its 3 x 8-pin PCIe power connectors see each of the 8-pin cables delivering up to 150W of power for a total of 450W of power delivery. This means that you can use any GeForce RTX 40 series GPU right up to the RTX 4080 and RTX 4090 graphics cards, but overclocking them and driving past 450W might require a different cable or a different cable adapter.

ASAKA recommends keeping your power cables away from heat sources, as it might cause damage to the cables, your graphics card, or more. However, I think that will be hard, considering the G-Nexus P24 adapter runs directly in front of your graphics card... you know, a major source of heat.

NVIDIA has had a bunch of giveaways throughout this year, starting from entry-level through to mid-range and up to its flagship GeForce RTX 4090... but now we're seeing some price drops on the MSRP of the RTX 4090 of close to $600.

The news is coming from NVIDIA's new DOOM 30th Anniversary sweepstakes competition, where you can win a custom NVIDIA GeForce RTX 4090 graphics card celebrating DOOM's incredible 30th anniversary. But now we're seeing that the last two competitions that NVIDIA held have the MSRP of the RTX 4090 and its custom backplate sitting at $1389.

Why is this significant? Well, earlier this year, NVIDIA held a competition in February 2023 where the MSRP of the giveaway of the GeForce RTX 4090 with a custom Atomic Heart backplate had an MSRP of $1979. Later in the year, the MSRP of the GeForce RTX 4090 Founders Edition was $1599, and then in November and now December, the MSRP has fallen to $1389.

NVIDIA's flagship GeForce RTX 4090 has been custom-modded by a Redditor, taking the original cooler off and replacing it with a triple-fan Noctua cooling solution, and it looks awesome. Check it out:

Redditor "zaME19" posted on the Small Form Factor PC subreddit, modding his Inno3D GeForce RTX 4090, which features a non-reference PCB and triple-fan cooler, where he 3D printed a completely new shroud for his RTX 4090. The shroud features a matte black design with the exact mounting points for 3 x Noctua NF-A9 fans.

Each of the Noctua AF-A9 fans is 92mm with a PWM design and SS02 bearing mechanisms, with each of the fans capable of pushing out 78.9 m3/h airflow with a quiet operation of up to 22.8 dBa. They've got a max speed of 2000 RPM, operating as low as 400 RPM for that quiet, flagship GPU gaming goodness. The fans even come with anti-vibration pads which help dampen the noise even more.

AMD's next-gen RDNA 4 GPU architecture has popped up with recent patches for Linux again, with the company adding patches for their AMDGPU LLVM shader compiler.

Previously, AMD didn't push into Linux as hard as their competitors like Intel and NVIDIA, with the new patches not offering anything overly exciting, but instead, it's just traditional instructions support, updating existing hardware registers, and a bunch of pull requests for "GFX12".

We don't know anything more about RDNA 4 than we did yesterday, but what's interesting here is that AMD is forming its graphics driver for Linux with early support for its next-gen GPU architecture. Wccftech ponders that AMD is going to have increased competition to open-source alternatives including MESA's Vulkan driver. AMD could have improved support for next-gen architectures from here on out, even though RDNA 4 is most likely at least a year away.

GALAX has just launched its new GeForce RTX 4070 20th Anniversary Edition graphics card, with a sleek look due to its hidden power connector. Check it out:

The company is celebrating its 20th anniversary, with GALAX China launching its GeForce RTX 4090 20th Anniversary Edition -- but having to pull it from the Chinese market after US sanctions saw the RTX 4090 banned in multiple countries, including China -- and now it's here with the GeForce RTX 4070 20th Anniversary Edition card.

GALAX uses a beautiful cooler design, with two fans on the front -- and to the left -- and another fan on the back, with an all-white design that makes me feel tingly. It's only a dual-slot design which I love as well, with the RTX 4070 20th Anniversary Edition measuring just 45mm thick, and with its hidden 8-pin PCIe power connector under a magnetic cover, GALAX really makes this an appealing RTX 4070 for gamers.

Last week, we reported on the rumor that NVIDIA was planning to stop producing the entry-level GeForce RTX 3050 8GB graphics cards in favor of a cutdown GeForce RTX 3050 6GB refresh. This refresh is launching soon, too, sometime in January 2024, potentially around the same time as the new GeForce RTX 40 Series SUPER GPUs.

In addition to less VRAM capacity, the GeForce RTX 3050 6GB refresh will supposedly sport fewer CUDA Cores (a 20% reduction), a boost clock of 1700 MHz, and a power rating of 75W - which would mean that it would get all of its power via the PCIe slot.

It's an interesting move, and according to the latest slice of info from Taiwanese outlet Benchlife, The GeForce RTX 3050 6GB could launch with a price point of only $179 - a 28% reduction from the GeForce RTX 3050's launch MSRP of $249.

For those that need a little refresher, AMD's Fluid Motion Frames (or AFMF) is the company's driver-based version of Frame Generation - as seen in NVIDIA's DLSS 3 and AMD's FSR 3. This version of the technology is a potential game-changer because it allows for Radeon RX 7000 and Radeon RX 6000 GPU owners to enable frame generation in all DirectX 11 and 12 titles. That's right, per-game support isn't required.

Still in its pre-release 'AMD Software: Adrenalin Edition Preview Driver for AMD Fluid Motion Frames' form, AMD quietly dropped a new update late last week. Promising improvements to "stutter and pacing" and "general improvements to the stability" of AFMF technology bring the tech one step closer to full release.

Though as to when that will be is anyone's guess. The good news is that the update aligns the Preview Driver with the latest AMD Software: Adrenalin Edition 23.12.1 release - so if you're keen to check it out, now's the time.

Acer is the second system builder that we know of to have its own GeForce RTX 4090, after LENOVO and its RTX 4090 LEGION, with the new Acer RTX 4090 Predator being independently tested by our friends at KitGuru.

The company is now shipping its new GeForce RTX 4090 Predator graphics card with an interesting hybrid-style cooler inside of its high-end Predator Orion X Gaming PCs. The first review unit has reached KitGuru, who has pulled apart the RTX 4090 Predator graphics card inside of it. The card doesn't actually have a name, but given that it's inside of a Predator gaming PC, it should be the RTX 4090 Predator, right?

Acer's custom GeForce RTX 4090 Predator graphics card is a triple-slot beast, that is almost -- but not quite as -- big as NVIDIA's in-house GeForce RTX 4090 Founders Edition. It weighs slightly less as well, at 1.8kg for the RTX 4090 Predator compared to 2.19kg for the RTX 4090 FE. Acer uses a matte black plastic shroud with embedded RGB accent LEDs that can be controlled through Acer's in-house Predator software.

Intel's Arc GPUs just got a new graphics driver which again brings in some impressive performance uplifts.

Driver version 31.0.101.5074 offers a rather tasty boost of up to 34% for Train Sim World 3 (at Full HD resolution, Ultra settings) and Arc graphics card owners are getting 23% more FPS in Battlefield V at those same 1080p settings.

Here's a full roundup of the frame rate uplifts provided by the new Arc graphics driver:

NVIDIA is working on a new workstation GPU that could be made specifically for China, with a tease of its new RTX 5880 Ada workstation GPU that has been spotted in some new NVIDIA RTX Enterprise drivers.

The new NVIDIA RTX 5880 Ada workstation GPU was spotted inside of the most recent production branch of the NVIDIA RTX Enterprise driver, which is a new workstation card that has been tweaked to comply with the US export restrictions, something that allows NVIDIA to get AD102 GPUs into China. It will join the coming AD102-250 GPU-based GeForce RTX 4090 D, which has also been tweaked to get into China.

US export restrictions first saw the NVIDIA RTX 6000 Ada workstation GPU and high-end gaming-focused RTX 4090 banned in China back in November 2023, with NVIDIA working on multiple variants of its workstation and gaming GPUs to comply with the ever-changing landscape of GPUs being banned in China (and other countries around the world).

NVIDIA is reportedly working on a refreshed GeForce RTX 4090 graphics card that has been customized for the Chinese market, with the upcoming GeForce RTX 4090 D... the "D" stands for Dragon because it's the Year of the Dragon in 2024 for the Chinese New Year.

We've heard that NVIDIA will be using a tweaked AD102-250 GPU for the RTX 4090 D, which will feature less CUDA cores than the AD102-300 GPU that powers the RTX 4090. The new RTX 4090 D will also have a lowered 425W TDP down from the RTX 4090 and its 450W TDP, but new information is here about the GPU and power consumption: the RTX 4090 D will NOT support overclocking.

NVIDIA's new GeForce RTX 4090 D being locked away from overclocking makes sense, given that it needs to not go over the US government and its constantly updated GPU restrictions in China, so these limitations need to happen. CUDA core counts, GPU clock speeds, TDP, and overclocking support for the RTX 4090 D will be what sets it apart from the RTX 4090.

The mainstream and affordable GeForce RTX 3050 8GB was the final Ampere GPU to hit the scene, with the card making its debut in January 2022 - so it's still relatively fresh in the world of discrete graphics cards. According to information from the Board Channels forum in China, NVIDIA is set to scrap the GeForce RTX 3050 8GB altogether and replace it with a cutdown 6GB model.

Rumors indicate that the new GeForce RTX 3050 will use the GA107 GPU from the mobile version and arrive with 20% fewer CUDA Cores (2048 versus 2560) and a boost clock of 1700MHz.

It's a pretty drastic cut, and the reason might come down to the new 75W power rating (compared to 130W for the existing model) - which would mean that the new desktop GeForce RTX 3050 6GB will get all of its power from the PCI Express slot. It's a strange move to replace the least powerful GPU in the current GeForce RTX lineup with a worse card.

The GeForce RTX 4090 is banned from sale in China thanks to new U.S. Government restrictions imposed to curb the country's advances in AI. And since that news dropped, it's been a roller coaster ride for the world's most powerful gaming GPU - with global prices increasing dramatically and rumors that NVIDIA plans to launch a China-exclusive GeForce RTX 4090D in January 2024. The Year of the Dragon, hence the 'D'.

The RTX 4090D uses the AD102-300 GPU, a cutdown version of the AD102-250 GPU in the existing GeForce RTX 4090 that will adhere to the new restrictions imposed on AI hardware exported to the region. The latest info from well-known NVIDIA leaker MEGAsizeGPU over on X/Twitter claims that the GeForce RTX 4090D will feature the same Boost Clock speed as the existing RTX 4090 (2520 MHz) but a higher Base Clock speed (2280 MHz).

Additional info from other sources claims that the RTX 4090D will also feature the same 24GB of GDDR6X memory, though further details like memory speeds and TDP are still being finalized. It's believed that NVIDIA will lower the CUDA Core count to fall within the restrictions that measure AI performance.

AMD's latest Adrenalin Edition 23.12.1 driver release for Radeon GPUs is a big one in that there's a lot more going on than simply adding day-one support for one of the last big PC game releases for 2023 - Ubisoft's Avatar: Frontiers of Pandora. The most immediate changes come with a redesign of the UI for the Home and Gaming tab to better integrate HYPR-RX.

HYPR-RX combines Anti-Lag+, Radeon Boost, and Radeon Super Resolution into a single-click option to make access to these technologies easier. However, due to issues with latency reduction interfering with game code, Anti-Lag+ is no longer a part of the equation, and there's no word on when it'll be coming back. Anyway, back to the UI redesign.

AMD's release notes state that HYPR-RX profiles can be enabled on a per-game basis on the Home Tab - a great addition as it brings AMD's tech to the forefront of the UI. Switching over the Gaming tab, users will be able to learn about each HYPR-RX feature and have direct control over any potential settings. It doesn't stop there, as Adrenalin Edition 23.12.1 introduces the brand-new AMD HYPR-RX Eco mode.

The mid-range GeForce RTX 4060 Ti comes in two flavors - the more common 8GB variant and one with 16GB of GDDR6 memory on a similar 128-bit bus. With a single 8-pin power connector and a TGP of 165W, the GeForce RTX 4060 Ti is one of the most power-efficient GPUs in the GeForce RTX 40 Series. This makes it the perfect candidate for smaller form factor models, cards designed to take up minimal space inside a build.

Enter the new GeForce RTX 4060 Ti Musou MAX or "Unchallenged" from Chinese NVIDIA partner GALAX, which features an ultra-thin single-slot form factor. I can't remember the last time I saw a single-slot graphics card capable of running the latest games, so this is pretty exciting.

With dimensions of 267x111x20 mm, the GALAX GeForce RTX 4060 Ti MAX includes a vapor chamber cooling system with a copper heatsink and a blower-style fan. It's a unique design because it marks the first GPU release for the GeForce RTX 40 Series to use a single-slot form factor.

A new report from Jon Peddie Research presents some great news for the PC gaming market: GPU shipments from AIBs for the desktop rigs have increased by 37.4% in the last quarter. When you factor in all desktop PCs and not those custom-built for gaming or sold as-is, GPU shipments increased 16.8% in Q3 2023 compared to the last quarter.

All three big players in the GPU space benefited from recovery, and the report proudly proclaims that the post-COVID slump for PC gaming hardware "is over." However, AMD and the Radeon line-up look to have benefited the most as AMD's shipments increased by 36.6%, NVIDIA's by 25.2%, and Intel's by 10.4%.

This sees AMD's overall market share increase by 2.4% compared to the last quarter, with NVIDIA's growing by 1.3%. The report notes that iGPUs and APUs account for the shift, and part of that will undoubtedly be due to the rise in popularity of PC gaming handhelds like Steam Deck and ROG Ally.