Graphics Cards - Page 12

NVIDIA's Smooth Motion technology was released earlier this year, a driver-based Frame Generation solution similar to AMD's Fluid Motion Frames in that it doesn't require game integration, like DLSS. Initially, this new feature was exclusive to the latest GeForce RTX 50 Series of graphics cards, but that's changing in an upcoming driver release.

The feature was spotted in the latest GeForce 590.26 Preview Driver for developers, which includes support for enabling Smooth Motion on GeForce RTX 40 Series graphics cards. Although Smooth Motion doesn't require game integration like DLSS Frame Generation, it still utilizes the AI hardware in GeForce RTX GPUs for generating its additional frame.

Smooth Motion, like Frame Generation, is all about improving the performance of a game by generating a new frame in between traditionally rendered frames. This improves the overall smoothness of the presentation, giving the impression and effect of doubling or substantially boosting performance.

ASUS has revealed the most expensive graphics card ever made, with the introduction of its new ROG Astral GeForce RTX 5090 Real Gold Edition, and it's worth as much as a house at over $500,000.

The new ASUS ROG Astral GeForce RTX 5090 Real Gold Edition was revealed at the recent Bilibili World 2025 event, with some pictures shared by @realVictor on X of the new ROG Astral GeForce RTX 5090 Real Gold Edition that is made purely of real gold. Not just up to 6.5g of gold on the recent ROG Astral Dhahab Edition graphics cards made for the Middle Eastern market, but real, actual gold, and a real, actual price of $500,000.

ASUS is using 5kg of real gold on its custom ROG Astral GeForce RTX 5090 Real Gold Edition graphics card, and with 1kg of gold selling for around $100,000... that means that this new custom graphics card is worth an insane $500,000 in just gold alone, with another few thousand dollars for the GeForce RTX 5090 that it's been dipped in.

NVIDIA has announced another giveaway, this time it's a custom Cyberpunk 2077-themed GeForce RTX 5090 Founders Edition graphics card, and you can enter by commenting on the X post below.

The company announced a couple of days ago it was giving away a custom Superman-themed GeForce RTX 5090 Founders Edition signed by Jim Lee, with the giveaways a part of the GeForce RTX Summer event. The Cyberpunk 2077-themed RTX 5090 isn't hand-painted like the Superman-themed RTX 5090, but it still looks slick.

There's a full yellow sticker wrap on the card, but it would've been real nice if it was a full custom job like previous special editions that the company has used on Cyberpunk 2077-themed graphics cards over the years.

COLORFUL is cooking up a new GeForce RTX 50 series iGame Ultra Duo White SSD GPU, with dual M.2 SSD slots on the back, letting you expand your SSD storage using your graphics card.

Chinese media noticed the new graphics card at the recent Bilibili World 2025 event, with what appears to be COLORFUL's exciting new iGame Ultra Duo White SSD, with what should be a GB206 or GB207-based GeForce RTX 50 series GPU with dual M.2 SSD slots on the back.

The PCB itself looks quite short, with mounting holes on the back of the GPU heatsink sitting between the SSD and I/O bracket without a full slot required. The upcoming COLORFUL iGame Ultra Duo White SSD GPU uses just 8 PCIe lanes, with two additional SSDs using the PCIe bifurcation mechanism, allowing multiple devices to be used on the same interface. The GPU uses 8 lanes, with each SSD enjoying 4 dedicated lanes.

GALAX's flagship GeForce RTX 5090D HOF OC Lab XOC graphics card is an elusive GPU that has just hit eBay with a starting price of $5500... check it out:

The new ultra-flagship GALAX RTX 5090D HOF OC Lab XOC graphics card is one of the fastest consumer GPUs on the planet, with dual 16-pin PCIe power connectors and an insane 2000W BIOS. It has broken multiple world records in the last few weeks since its introduction, with more details on that in the links below.

GALAX's custom RTX 5090D HOF OC Lab XOC graphics card has a powerful 38-phase PWM, 26 of which are dedicated to the GPU, with a 12-layer PCB and two 12VHPWR power connectors. The dual 16-pin power connectors and its 2000W BIOS are designed for LN2 cooling, that allow for peak power consumption of over 1000W, far up and above the 600W limit that NVIDIA imposes on its AIB partners for the RTX 5090.

AMD has been rumored for quite a while that it was working on a multi-chiplet gaming GPU, with a new patent fueling these rumors in the battle of GPUs against NVIDIA.

In a new video shared by coreteks, we're hearing that AMD has recently secured a new multi-chiplet GPU patent that could solve the latency issues using a 'Smart Switch'. MCM (Multi-Chiplet Module) isn't new, as they're used outside of the gaming GPU market, but there are limitations of monolithic GPU designs that MCM can solve.

AMD used an MCM design approach for its Instinct MI200 series AI accelerators, stacking multiple chiplets onto a single package, including the GPCs (Graphics Processing Cores), HBM memory stack, and the I/O die. However, AMD used a different approach with its Instinct MI350 series AI chips, and this could be the foundation of its multi-chiplet-based consumer gaming GPUs according to coreteks in his new video.

The new PNY GeForce RTX 5050 Single Fan Graphics Card is one of the smallest desktop RTX GPUs, measuring just 146.7mm in length. The entry-level GPU from NVIDIA targets 1080p gaming performance, with DLSS 4, while also bringing the latest generation of hardware encoders and decoders to the compact home theater and content creator PC market.

The dual-slot single-fan PNY GeForce RTX 5050, alongside the company's GeForce RTX 5060 variant, is small enough to fit inside even the most compact case, thanks to its 130W power rating, heatsink, and optimized fan, ensuring it runs cool even during intense gaming sessions. Essentially, it resembles a dual-fan GPU cut in half, with the single 8-pin power connector positioned closer to the PCIe bracket than to the rear.

Port-wise, it's your typical GeForce RTX 50 Series GPU, featuring three DisplayPort 2.1 ports and a single HDMI 2.1 port. One of the most significant differences between NVIDIA's entry-level GeForce RTX 5050 and its counterparts is that it features the fewest CUDA Cores of the entire line-up. It has 8GB of older GDDR6 memory, whereas all other models utilize the faster and newer GDDR7 memory.

ASUS is celebrating 30 years in the GPU business, but now it's planning a new "special" GeForce RTX 50 series GPU, which could be the flagship ROG MATRIX GeForce RTX 5090 graphics card.

In a new special website where ASUS explains the terms and conditions of its 30 years in the GPU business, the company details its prize pool, explaining: "Join us in celebrating three decades of innovation and excellence in graphics card technology. Play your way through thrilling challenges in our commemorative browser game to earn points and for a chance to win incredible prizes, including a coveted ASUS GeForce RTX™ 50 series graphics card and much more".

In the Grand Prize section, ASUS teases "something BIG is coming soon" and will be a GeForce RTX 50 series GPU of some form, and it'll be released on August 19. ASUS says "these prizes include GeForce RTX 50 series graphics cards and power supply units, as well as a top-secret grand prize we'll announce on August 19".

The new Superman movie is here and NVIDIA is celebrating in the best way possible with a new giveaway, a one-of-a-kind Superman-themed GeForce RTX 5090 graphics card that has been signed by the legendary Jim Lee.

The official NVIDIA GeForce account on X posted about the giveaway, where you can win the one-of-a-kind Superman-themed GeForce RTX 5090 Founders Edition that has been hand-painted and signed by Jim Lee. Lee used acrylic paint, featuring Superman and his loyal companion Krypto. On the card's front face, Lee personally signed it, too.

Lee is the President, Publisher, and Chief Creative Officer of DC Comics, with NVIDIA's new Superman-themed GeForce RTX 5090 Founders Edition graphics card giveaway representing a huge chance for Superman fans. As for the movie, Superman hits theaters on July 11, directed by James Gunn, who was behind the camera for the Guardians of the Galaxy trilogy when he was at Marvel.

NVIDIA currently uses SK hynix and Samsung GDDR7 memory chips for its GeForce RTX 50 series GPUs, but Micron is about to step into the GDDR7 memory chip supply chain.

In a new report from Benchlife, supply chain sources have said that in addition to SK hynix and Samsung GDDR7 memory chips, NVIDIA has added Micron's new GDDR7 memory for its RTX 50 series. This means that there are now 3 versions of GDDR7 memory to choose from moving forward.

Micron announced its GDDR7 memory chips in a press release in 2024, announcing that it had started sampling GDDR7 memory chips. As it stands, Micron's new GDDR7 memory is available in two different versions: 28Gbps and 32Gbps, on its 1β process.

There have been some interesting additions for the latest version of AIDA64 in terms of AMD hardware, hinting at the prospect of a cheaper RDNA 4 GPU that could be inbound.

VideoCardz flagged up that the notes for the latest stable release of AIDA64 include not just the AMD Radeon RX 9060 XT, but also the vanilla RX 9060.

That is, of course, a GPU which AMD is yet to confirm, but its appearance here is a strong suggestion that it might be due before too long (with any luck). It's worth noting that NVIDIA's RTX 5050 and 5060 are included with the new GPUs supported here. Oh, and there's Intel's BMG-G31 Battlemage graphics card, too.

AOOSTAR has launched its new XG76XT external graphics card enclosure, which features a built-in RDNA 3-based AMD Radeon RX 7600 XT.

AOOSTAR's new XG76XT external GPU enclosure is listed on Chinese retailer JD for 3399 yuan (which converts to around $470 USD) with a regular retail price of 3499 yuan. The new external GPU solution is perfect for laptop or compact PC owners that want more graphics performance, without the bulk of a desktop PC.

The new XG76XT external GPU uses a full desktop-class Radeon RX 7600 XT, packing 32 CUs of RDNA 3 GPU power, and 16GB of GDDR6 memory on a 128-bit memory interface. The Radeon RX 7600 XT graphics card supports 190W TGP in desktop configurations, while AOOSTAR's new XG76XT external GPU is limited to 150W. The AOOSTAR XG76XT at 150W will beat mobile versions like the RX 7600M XT, which is in other external GPUs, which is usually limited to 120W.

Leave it to modders to make quieter versions of the NVIDIA GeForce RTX 5090 and RTX 5080 graphics cards, with some rather awesome-looking custom Blackwell cards. Check them out below:

The better-looking of the two -- at least to me -- is from "Ultimate Shrouds", which is in the video above. This mod uses the MSI GeForce RTX 5090 VENTUS graphics cards which can now sport an incredible 4 or even 5 fans.

We have a custom 3D-printed shroud that holds 2 x Phanteks T30 fans, and 2 x Noctua NF-A6x25 PWM fans, but this mod can also support a third T30 fan on the back.

GALAX's flagship GeForce RTX 5090 D graphics card has set multiple new overclocking world records, seeing its GPU clocks hit a huge 3750MHz and its 32GB of GDDR7 memory clocked at an insane 36Gbps.

Team OGS is an overclocking team that has pushed the GALAX RTX 5090 D to its limits, using the GPU inside of a system powered by an Intel Core i9-14900KF processor inside of an ASUS ROG Maximus Z790 APEX motherboard.

Team OGS was running an XOC BIOS that pushes the maximum power limit up to a bonkers 2000W, with the GALAX RTX 5090 D itself featuring dual 12V-2x6 power connectors that are capable of delivering 600W per connector, for up to 1200W of power flowing into the card, leaving a bunch of OC headroom within the BIOS.

NVIDIA's flagship GeForce RTX 5090 loses over 25% of its performance in professional video editing workloads when it isn't fed its full PCIe bandwidth in new testing.

The folks over at Puget Systems have tested the GeForce RTX 5090 to see what performance impact there is when the card is under limited PCIe bandwidth, as there are situations in particular systems when this is a major issue. Some users may have the RTX 5090 inside of a system that has another add-in card that is using those precious PCIe lanes, and that reduces the performance of the GPU.

Puget Systems started the testing with After Effects where the RTX 5090 loses a large chunk of performance when the GPU drops from PCIe 5.0 x16 to PCIe 3.0 x4 with over 10% performance loss, while the same performance loss is noticed with DaVinci Resolve where over 20% performance is lost when the GPU is running in PCIe 3.0 x4 with performance also being lost when the number of lanes is dropped from x16 and x4 in the same Gen5 slot. This means multiple AICs on your motherboard can definitely have a huge performance impact on the RTX 5090.

The Steam Hardware Survey results for June 2025 are in, and they show that NVIDIA's full line-up of GeForce RTX 50 Series GPUs is quickly gaining market share. The most popular model is the GeForce RTX 5070, which has cracked the Top 25 discrete GPUs among PC gamers.

The RTX 5070 is NVIDIA's latest mid-range offering, succeeding the popular GeForce RTX 4070. Although the RTX 5070 made significant gains in June, the new GeForce RTX 5060 and GeForce RTX 5060 Laptop GPUs are both storming up the charts, making their debut after launching in late May - an impressive achievement.

Following the GeForce RTX 5070, the second most popular GPU in the 50 Series, so far, is the GeForce RTX 5080, followed by the GeForce RTX 5070 Ti in third place. In fact, with the RTX 5060, RTX 5060 Ti, and the flagship RTX 5090 also in the list, the entire lineup is accounted for - barring the brand-new GeForce RTX 5050, which launched this week.

The launch of the desktop GeForce RTX 5050, starting from $249, is the latest NVIDIA launch where review samples were not sent to the media and press ahead of time. At TweakTown, we're in the process of arranging a sample; however, the first RTX 5050 review has already appeared online, courtesy of the Korean outlet Quasar Zone.

The review covers the custom COLORFUL GeForce RTX 5050 ULTRA OC White model, paired with an AMD Ryzen 7 9800X3D processor and 32GB of DDR5 memory. With 15 games tested at 1080p, the review compares the RTX 5050's performance to a wide range of entry-level and mainstream GPUs.

Across the 15 titles, the GeForce RTX 5050's performance is within 6.2% of the RTX 4060, while it falls short (2.6%) of the similarly priced Intel Arc B580 'Battlemage' GPU. It's worth highlighting that the Intel Arc B580, although slightly faster than the RTX 5050, also ships with 50% more VRAM, featuring 12GB of faster GDDR6 memory.

File this one under 'ludicrous experiments' (as well as improbable ones), but it turns out that if your GPU ever appears in the list of USB media that can be ejected from the taskbar in Windows, you really shouldn't try to do this.

But that's exactly what a foolhardy Redditor did - curious as to what the result would be - and hey, what do you know, but it ended up breaking the system (temporarily).

The first question you're likely to have is: How on earth did this happen in the first place? We've never had the option to eject our GPU like it was a USB stick, either, and apparently this is only served up in a virtual machine.

HighPoint has just announced the launch of their new RocketStor 8631CW and RocketStor 8531AW Gen5 x16 eGPU enclosures, which have been purpose-built to handle the high-end performance of the GeForce RTX 5090, pumping up to 600W of power into the eGPU enclosure.

The new HighPoint RocketStor 8631CW and RocketStor 8531AW Gen5 x16 eGPU enclosures are ready for space-constrained workstations and server environments, perfect for professionals who are space-limited but still need a ton of GPU power.

HighPoint's new eGPU enclosures provide a dedicated external PCIe Gen5 x16 expansion pathway, which enables the full utilization of cutting-edge GPUs like NVIDIA's flagship GeForce RTX 5090, all without compromising the thermal performance, PCIe bandwidth, or system integrity of compact workstation and server platforms.

NVIDIA's new GeForce RTX 5090 D v2 is reportedly the name of its new China-specific GPU, according to a new post on X by leaker "MEGAsizeGPU".

In another post, HKEPC reports that NVIDIA is preparing its new GeForce RTX 5090 D v2 for an August launch, something we heard about in previous reports. This means that gamers in China will only have another month before they can secure themselves the new (tweaked) flagship GeForce RTX 5090 D v2 graphics card.

Inside, the NVIDIA GeForce RTX 5090 D v2 has the same GPU core count as the RTX 5090D with 21760 CUDA cores and the tweaked GB202-240-K-A1 GPU. The VRAM has been scaled down to 24GB (down from 32GB on the RTX 5090) on a 384-bit memory bus clocked at 28Gbps and the same 575W TDP but leaks tease "and there's a surprise".