Graphics Cards - Page 85

NVIDIA doesn't update its GeForce Experience app often, so the latest version - GeForce Experience 3.27 - is big. For those who like to capture high-quality game footage, you'll be pleased to know that it fixes a HDR bug, so 8K 60FPS HDR recording with a high-end GeForce RTX 40 Series GPU will no longer look overexposed.

Interestingly, the release notes mention that this new version "adds support for GeForce RTX 40 Series Laptops - from ShadowPlay & Optimal Game Settings to Max-Q technologies like DLSS 3, Whisper Mode, and Battery Boost 2.0." Arriving almost a year after these laptop chips hit the market, it will be interesting to see if it makes a noticeable difference to things like battery life and the quality of Max-Q tech.

The GeForce Experience app from NVIDIA is also great for optimizing game settings for performance, and version 3.27 adds "optimal settings" support to 50 games for GeForce RTX GPU owners. It's a massive list that includes everything from Marvel's Spider-Man Remastered to Overwatch 2 and Warhammer 40K: Darktide.

3DMark synthetic GPU benchmarks have been a staple part of the review process here at TweakTown for a long time, and over the years, the two big standout tests have been the DirectX 11-based FireStrike and the DirectX 12-based TimeSpy. For those buying new graphics cards, firing up 3DMark TimeSpy right after installation to check out the score has become a right of passage.

Other benchmarks that cover ray tracing (like Port Royal) and other high-end GPU features have appeared. Still, as 3DMark enters its 25th anniversary, it is set to debut a new staple benchmark called Steel Nomad, just in time for NVIDIA's new GeForce RTX 40 Series SUPER graphics cards that will launch early next year.

3DMark Steel Nomad is a forward-thinking benchmark that will work across multiple platforms, APIs, and operating systems. Windows users will use the modern and updated DirectX 12 API; for macOS users, it will use Metal; and for Android and Linux systems, it will use Vulkan.

The Steam Hardware and Software Survey results for October 2023 are in, and it's very good news for NVIDIA - with its share of the total GPU market increasing by an impressive 5% compared to the September 2023 results. It's now 82.2%, with second-place AMD's market share rapidly declining from around 15% in September to 11.5% in October 2023.

For NVIDIA's current GeForce RTX 40 Series of GPUs, the Ada Lovelace generation is starting to make some major moves - with the most popular model in the lineup being the GeForce RTX 4070. It's sitting just outside the top ten, too, so odds are it'll crack the full ten GPUs out in the wild before the year ends. Especially if the recent discounts that brought the GeForce RTX 4070 price down to USD $549 hold.

The GeForce RTX 4070's popularity makes sense; it's one of the best value GPUs currently available - delivering excellent 1440p performance with next-level efficiency. The GeForce RTX 4070's position in the top GPUs is closely followed by the GeForce RTX 4060 and the GeForce RTX 4060 Ti.

AMD has just announced its latest game bundle, which will see Ryzen 7000 and Radeon RX 7000 series products purchased, including a free copy of Avatar: Frontiers of Pandora.

The Avatar: Frontiers of Panodra game bundle has already started, running through to December 30 at 11:59 PM ET, with all product codes needing to be redeemed by January 27, 2024. If not, the coupon will expire, and you'll no longer be able to download or redeem your free copy of the game.

As for which Ryzen 7000 series and Radeon RX 7000 series products, a free copy of Avatar: Frontiers of Pandora is included with the following processors: Ryzen 9 7950X3D, Ryzen 9 7950X, Ryzen 9 7945HX3D, Ryzen 9 7900X3D, Ryzen 9 7900X, Ryzen 9 7900 and the Ryzen 7 7800X3D, meanwhile, on the GPU side of things, you'll get Avatar: Frontiers of Pandora for free with the following GPUs: the Radeon RX 7900 XTX, Radeon RX 7900 XT, Radeon RX 7900 GRE, Radeon RX 7800 XT, Radeon RX 7700 XT, and the Radeon RX 7600.

AMD's older Polaris and Vega GPUs aren't having support withdrawn, but that process is definitely underway.

That's according to a report from Anandtech, which observes that AMD recently ceased support for these graphics cards in its Vulkan driver for Linux - and the company is stepping back with Windows drivers, too.

This doesn't mean support has been pulled totally, but AMD will only deliver critical (security) updates for Polaris and Vega, and functionality updates will be provided where they are available.

COLORFUL teased its new custom GeForce RTX 40 series iGame Ultra Z graphics cards are starting to launch, with the new GeForce RTX 4060 Ti iGame Ultra Z graphics card available in China.

The first of which is the new GeForce RTX 4060 Ti iGame Ultra Z graphics card, which is the second-best of the Ultra Z series cards from the company. The new iGame Ultra Z cards feature the more common 8-pin PCIe power connector that is kept securely in place by a magnetic backplate.

So, if you need to get to your COLORFUL GeForce RTX 4060 Ti iGame Ultra Z graphics card, all you'll have to do is pull the magnetic backplate off, disconnect the 8-pin PCIe power connector, and you're good to go. Another benefit from COLORFUL is that they're using the traditional 8-pin PCIe power connector, so there's no special 12VHPWR power connector that could go awry with bending over time or overheating issues over the years of use inside of your gaming system.

NVIDIA's upcoming refreshed GeForce RTX 4070 Ti SUPER graphics card has been teased in more "official" form with some leaked retail packaging renders from "MEGAsizeGPU" on Twitter.

The new retail packaging of the NVIDIA GeForce RTX 4070 Ti SUPER graphics card doesn't show us anything we didn't know already, but just more confirmation of the box, I guess. Inside, the RTX 4070 SUPER will feature either the AD102-175 or AD103-275 GPUs with 8488 CUDA cores while packing an increased 16GB of GDDR6X -- up from the 12GB on the RTX 4070, RTX 4070 Ti, and RTX 4070 SUPER graphics cards.

One thing worth mentioning is that the "SUPER" logo seems to be getting a facelift, with a simple design that better aligns itself with the "NVIDIA Sans" font, which is used across the "GeForce" branding and the model SKU numbers. What we can see here with the RTX 4070 Ti SUPER retail branding is that the "SUPER" branding would be placed next to the "Ti" designation.

NVIDIA's newly refreshed GeForce RTX 40 SUPER series of graphics cards have been teased for a while now, with a leaker teasing they should be revealed at CES 2024 in January 2024.

We've had leaks from "kopite7kimi" for years now, with someone asking when they -- the GeForce RTX 40 SUPER cards -- might be coming out, to which the leaker responded simply with "CES." CES -- or the Consumer Electronics Show -- takes place each year in early January and has been a shadow of its former self since the pandemic.

NVIDIA won't have a gigantic presence at the show, but the RTX 40 SUPER graphics cards aren't a next-gen GeForce RTX series but a refresh of their already mega-successful RTX 40 series graphics cards. NVIDIA will be showing off multiple things: new partner custom GPU designs, new technologies -- DLSS, Reflex, etc -- and, of course, AI.

Nintendo has yet to formally announce or confirm any sort of detail when it comes to the upcoming Nintendo Switch 2. Still, based on rumors, leaks, and insider info, there's a lot of info out there on just what sort of hardware will power Nintendo's next-generation handheld console.

And with that, the tech gurus at Digital Foundry have put the potential Ampere-based NVIDIA hardware inside to the test.

With games like Death Stranding, Cyberpunk 2077, A Plague Tale Requiem, Control (with ray-traced reflections), and the Unreal Engine 5 version of Fortnite put to the test, it's an interesting look at what the upcoming Nintendo Switch might be capable of - and just how much of a difference DLSS will make when it comes to presentation and performance.

The AI industry is absolutely booming worldwide, with the likes of NVIDIA sold out of its Hopper H100 AI GPUs for the next 6 to 9 months at the least, and now we're hearing HBM3 memory makers SK hynix and Samsung are sold out of HBM3 memory until 2025.

The news is coming from ZDnet in Korea, which reports that because of the non-stop need for AI GPUs that use super-fast HBM3 memory, Samsung and SK hynix are celebrating. Why? Because they're the only two companies making HBM3, it's completely sold out to AMD and NVIDIA through the entire of 2024 and into 2025.

HBM3 memory is one of the keys of high-end and next-gen AI GPUs because of the huge transfer speeds, faster bandwidth, increased memory capacities, and lower power consumption. SK hynix is making most of the HBM memory, but it might not be for long because Samsung is trying its best to catch up with its new HBM3e "Shinebolt," recently announced.

NVIDIA has just pushed out a new hotfix driver, which tackles the issues with Alan Wake 2 and its stability and performance degradation in the form of the GeForce Game Ready 546.08 Hotfix drivers.

We see hotfix drivers over time, with NVIDIA engineers finding an issue in its drivers that is something important -- in this case, issues with Alan Wake 2, which shines on GeForce RTX series GPUs -- and so we get a quicker hotfix driver release. We will see a WHQL release in the coming days if you don't like using hotfix drivers.

The only issue in games that these new GeForce Game Ready 546.08 Hotfix drivers fix is Alan Wake 2, so if you're using a GeForce GPU and featuring any stability and performance issues, then you'll want to use these hotfix drivers. You can download the new GeForce Game Ready 546.08 Hotfix drivers here (669MB download).

AMD unveiled its new Radeon RX 7900M mobile GPU a few weeks ago, with the Navi 31-powered GPU now being tested in 3DMark TimeSpy, where it is 4% faster than the competing NVIDIA GeForce RTX 4080 Laptop GPU.

The new AMD Radeon RX 7900M uses the Navi 31 GPU with 4608 Stream Processors, with a Game Clock of 1825MHz and a Boost Clock of up to 2090MHz. It is being directed right at NVIDIA's GeForce RTX 4080 Laptop GPU, with the RX 7900M packing. 16GB of GDDR6 memory clocked at 18Gbps on a 256-bit memory bus, while the Total Graphics Power (TGP) sits at up to 180W.

AMD is really pushing the limits with 180W TGP and up to 200W using SmartShift Max technology, while NVIDIA's flagship GeForce RTX 4090 Laptop GPU with 16GB GDDR6 memory also has a TGP of 150W but can hit up to 175W with Dynamic Boost technology. AMD is pushing the power available to its RX 7900M, so I'm keen to see real performance-per-watt numbers from reviewers in the coming weeks.

GALAX teased its new GeForce RTX 4090 20th Anniversary Edition graphics card not too long ago, and now we're hearing some new information: it will be the first desktop graphics card to have an HDMI retimer chip, which allows for longer HDMI cables, something that's actually really handy.

Why is this useful? HDMI retimers have multiple functions and can be installed on not just graphics cards but the receiving end or repeater devices. They will eliminate noise and jitter from signals -- something that's needed for the much higher bandwidth that's being driven through HDMI 2.1, especially important with 4K 120FPS -- over longer HDMI cables.

Personally, one of my systems is plugged into a large 77-inch 4K 120Hz LG OLED TV and an HDMI 2.1 cable, and there's only a certain length I can go up to and maintain the best connection quality. Not that you'd buy this graphics card purely for the HDMI retimer, but it's an enthusiast GPU, so this small touch actually stands out from the rest of the RTX 4090 cards, which is why it's being pointed out. Nice touch, GALAX.

NVIDIA has its flagship GeForce RTX 4090 graphics card on the market with 24GB of GDDR6X memory, but there was a prototype GeForce RTX 4090 Ti made with a gigantic quad-slot design... and that I/O shield has just hit eBay.

We previously saw the GeForce RTX 4090 Ti prototype quad-slot cooler hit the Chinese black market, and now we've got the RTX 4090 Ti quad-slot I/O bracket. It features the regular 3 x DisplayPort 1.4 connectors and a single HDMI 2.1 port. VideoCardz reports that we're probably seeing the suppliers that made these RTX 4090 Ti quad-slot I/O brackets wanting to offload them because they're useless -- NVIDIA won't release the GeForce RTX 4090 Ti, it seems -- to make some money.

They're actually not badly priced at $25 with free shipping, making for quite the collector's item on the cheap: an unreleased graphics card -- and not just something mid-range, but what would've been a powerful GeForce RTX 4090 Ti -- and while it's only an I/O shield, it's also only $25. Not bad.

It seems the ban on NVIDIA's RTX 4090 graphics card being shipped to China is going to be in place in two weeks, or so further chatter from the GPU grapevine indicates.

According to Tom's Hardware, with the website citing the usual insiders ('people familiar with the matter') and its own delving into published SEC filings, the RTX 4090 ban will kick in come November 17.

This backs up a report from MyDrivers earlier this week which specified the same date, just a fortnight from today.

AMD could see its AI upscaling technology powering Samsung's next-gen Galaxy S24 smartphones, with new rumors teasing that AMD is working directly with Samsung and Qualcomm on getting FSR onto smartphones.

The new rumor is coming from Tech_Reve on Twitter, who says it's an exclusive: "AMD, Samsung, and Qualcomm have decided to jointly develop 'FidelityFX Super Resolution (FSR)' in order to compete with NVIDIA's DLSS, and it is anticipated that FSR technology will be implemented in Samsung's Galaxy alongside ray tracing in the future."

Apple has its in-house MetalFX upscaling technology on its iOS devices, NVIDIA DLSS is all over the PC and laptop scene as well as the future-gen Nintendo Super Switch (I seriously think it'll be called that; Nintendo would be missing out), and then AMD FSR technology entering the smartphone business makes sense. But remember, Qualcomm has its own upscaling technology with Snapdragon GSR. What's happening with that?

NVIDIA has pushed out its new GeForce Game Ready 546.01 WHQL drivers, which offer the best day-one gaming experience for multiple games, including Call of Duty: Modern Warfare III.

Not only do the new GeForce Game Ready 546.01 WHQL drivers provide the best support for Modern Warfare III, but also Desynched, Jusant, and RoboCop: Rogue City. There are some fixes for issues in previous drivers, with the new drivers tackling Halo Infinite, where NVIDIA has fixed increased loading times after updating to the previous 545.84 drivers, as well as the significant performance drops on Maxwell-based GPUs.

In the new 545 release drivers, we've got support for CUDA 12.3 as well as RTX Video Super Resolution v1.5 technology, and of course, the latest performance improvements, bug fixes, and driver enhancements.

AMD has released AMD Software: Adrenalin Edition 23.11.1, which adds launch support for Call of Duty: Modern Warfare 3, Like a Dragon Gaiden: The Man Who Erased His Name, The Invincible, and JX3 Ultimate.

The new driver also adds Radeon Boost support to Alan Wake 2, which is AMD's performance-boosting technology that dynamically adjusts the resolution based on how much action is currently on screen or when there's a lot of fast movement. Remedy's Alan Wake 2 is a deliciously slow-paced game, so we're not sure exactly how it'll work there - but it's a nice get nonetheless.

There's other good news is that several issues are getting fixed - including Counter Strike 2 performance on some Radeon GPUs and flickering when playing Total War: Pharaoh and Alan Wake 2. Plus, some very welcome AI and Machine Learning improvements. However, there's no word on when Anti-Lag+ will be reinstated. Or, a timeline for when Fluid Motion Frames is leaving beta.

NVIDIA's upcoming GeForce RTX 40 SUPER series graphics cards have filled the headlines for weeks now, with the latest leaks on them teasing TDP numbers for the RTX 4080 SUPER, RTX 4070 Ti SUPER, and finally, the RTX 4070 SUPER.

According to new leaks from kopite7kimi suggests that the two higher-end SUPER refreshes in the GeForce RTX 4080 SUPER and GeForce RTX 4070 Ti SUPER will have the same TDP as their non-SUPER counterparts of 320W and 285W, respectively. The GeForce RTX 4070 SUPER, on the other hand, will have a 220W TDP, 20W increase on the 200W of the RTX 4070 non-SUPER.

However, these power numbers would be for NVIDIA's upcoming GeForce RTX 40 SUPER Founders Edition graphics cards, with AIB partners releasing custom variants of the GeForce RTX 4080 SUPER, RTX 4070 Ti SUPER, and RTX 4070 SUPER graphics cards having higher TDP numbers.

Intel has made great progress with its Arc graphics card drivers since its Alchemist GPUs first emerged, and the latest release is no exception.

The new Arc & Iris graphics driver version 31.0.101.4952 - which it should be noted is a beta - brings in support for the following games:

As well as support for The Talos Principle 2, this gets a solid performance boost, with the new driver offering an (up to) 19% increase in average frames per second at 1440p resolution (on High settings).