Video Cards & GPUs - Page 6

Images of PowerColor Radeon RX 9070 GRE Red Devil and PowerColor Radeon RX 9070 GRE Reaper GPUs have made their way online (via Videocardz), pointing to an imminent release for what could be the third RDNA 4 GPU to hit the market.

Usually, the GRE or Great Radeon Edition cards hit the Chinese market first, and that's expected to be the case here. With both images showcasing retail packaging alongside the GPU designs (these are the exact sort of images that show up on retail listings), this third Radeon RX 9070 GPU will reportedly hit the market before the rumored June launch for the mainstream Radeon RX 9060 XT.

Regarding performance, the Radeon RX 9070 GRE will sit between the Radeon RX 9060 XT and the Radeon RX 9070 and will ship with 12GB of GDDR6 memory on a 192-bit bus. Previous reports indicated that the Radeon RX 9070 GRE launch was pushed back to Q4 2025. However, Videocardz's sources state this isn't the case.

NVIDIA's RTX 5000 graphics cards may finally be turning a corner with stock levels improving considerably, or at least that's the latest suggestion from the grapevine.

Moore's Law is Dead (MLID) aired this rumor in his latest episode of 'Broken Silicon' on YouTube (see above), which typically recaps (or expands on) existing recent leaks, but sometimes introduces fresh material to spice things up more.

In this case, MLID has heard new info from a couple of sources at retailers about how NVIDIA is now getting more Blackwell GPU supply through.

AMD's RX 9060 XT graphics cards might be going on sale as soon as May 18, somewhat earlier than the timeframe the rumor mill has previously forecast for the release date.

If you recall that previous buzz on the grapevine, it claimed that AMD was looking at a Computex reveal (the show starts on May 20) for the RX 9060 XT GPUs (8GB and 16GB models), followed by a release in June.

That may well have been true at the time it was reported, and indeed the new rumor - coming from the Board Channels over in China (via VideoCardz) - insists that AMD was indeed planning this. However, apparently Team Red has now brought forward the release to May 18.

Igor's Lab has just released a new in-depth report that uncovers a potential 'hotspot issue' affecting all GeForce RTX 50 Series GPUs, unrelated to the GPU chip or the latest GDDR7 memory. The hotspots have been recorded in the power delivery systems on the PCB, which Igor's Lab says could affect the longevity of affected graphics cards built using NVIDIA's reference designs.

Igor's Lab tested a range of GPUs from NVIDIA partners like Palit, PNY, MSI, and others. It describes the issue as "heat nests" located below the power supply that aren't often actively or otherwise cooled. And the issue seems to be worse on lower-tier cards like the GeForce RTX 5070, as the shorter and smaller PCB means these sections are more densely packed between the GPU and the display outputs.

Looking at a PNY GeForce RTX 5080 and RTX 5070, the former records a hotspot measuring 80.5 degrees Celsius, while the RTX 5070 reports a hotspot of 107.3 degrees Celsius. In the case of the RTX 5070, this is with a GPU temperature of 69.7 degrees.

NVIDIA notably didn't send out review samples of the RTX 5060 Ti 8GB version - when the new Blackwell GPUs were launched last week, only the 16GB flavor was provided to reviewers - and we can see why going by a new YouTube video.

This is Hardware Unboxed (as flagged by VideoCardz), which had to buy an RTX 5060 Ti 8GB graphics card to give it a thorough testing. And - spoiler alert, albeit this is nothing you won't guess - the verdict was not good. Not good at all.

It seems that 8GB really isn't enough of a video RAM pool these days, although gamers have been saying that since the RTX 4060 Ti, of course (clearly NVIDIA wasn't listening then, or now).

With NVIDIA launching the GeForce RTX 5060 Ti last week and announcing the GeForce RTX 5060 for May 2025, AMD's RDNA 4 response is set to be the Radeon RX 9060 XT. According to the latest reports and information from a known insider, Hoang Anh Phu (@AnhPhuH) The current expectation is that the GPU will be announced at Computex 2025 and launched on-shelf two weeks later.

Computex 2025 is happening a little earlier this year, from May 20 to 23 in Taipei, which would put the Radeon RX 9060 XT launch window in the first week of June. This lines up with AMD confirming that the Radeon RX 9060 XT would arrive sometime in Q2 2025; however, this is a little later than some expected.

AMD's RDNA 4 debut is still fresh in our minds. The flagship Radeon RX 9070 XT's March 6 launch delivered one of the company's best GPUs in years, thanks to the massive improvements to ray-tracing performance and the arrival of the new AI-powered FSR 4 upscaling. Plus, the aggressive $599 pricing (which, granted, only applies to a handful of models) meant that it quickly sold out in several locations.

As reported, NVIDIA recently released GeForce Game Ready 576.02 WHQL for the new GeForce RTX 5060 Ti. The driver also introduced many fixes and improvements addressing issues like random black screens and crashing. However, since its release, gamers began to report a weird GPU temperature monitoring bug with GeForce Game Ready 576.02. Potential incorrect readings could lead to the wrong clock speeds and voltages being applied after a PC wakes from sleep.

The good news is that NVIDIA and the GeForce team have quickly addressed this issue, especially with GeForce Hotfix Display Driver 576.15. As a hotfix or beta driver release, this driver update has gone through an "abbreviated QA process," so it's flagged as optional. However, with the GPU temperature monitoring bug and a few other GeForce RTX 50 Series fixes, the latest GeForce Hotfix Display Driver is worth installing if you are affected.

GeForce Hotfix Display Driver 576.15 is available via the NVIDIA Customer Care support site. Here are the release notes.

It seems Intel will be launching refreshed Arc Battlemage GPUs at Computex 2025 according to the latest rumors, boasting a larger 24GB of GDDR6 memory.

We have seen rumors that an upgraded Arc Battlemage GPU could be in the works with an updated memory configuration, doubling the Arc B850 and its 12GB of GDDR6 memory on a 192-bit memory bus, up to 24GB. But now Intel AIB partner SPARKLE has teased just that: a new 24GB Arc Battlemage is launching in "May-June".

In a post on Chinese social media outlet Bilibili, SPARKLE China replied to someone asking if we'd see a 24GB variant coming out, to which they said "how do you know" and then replied, saying: "the original plan was May-June". Once this comment hit the internet, SPARKLE walked back its comment, and issued an official statement regarding the existence of the purported GPU.

AMD is working on a new RDNA 4-based workstation GPU that will feature 32GB of GDDR6 memory, which should be released as the Radeon PRO W9000 series.

In a new post on X by leaker @AnhPhuH, we're hearing that the new "Navi 48 XTW" GPU will feature 32GB of VRAM, and will be part of a new PRO series of workstation GPUs. AMD launched its RDNA 4-powered Radeon RX 9070 series GPUs recently, with its flagship Radeon RX 9070 XT featuring 16GB of GDDR6 memory.

The purported Navi 48 XTW GPU with 32GB of GDDR6 will be its highest-spec configuration in terms of VRAM, with consumer GPUs left with only 16GB. For AI developers, the additional 16GB (32GB total) will be a big win, but it's nothing we haven't seen before.

Intel is looking to hire someone that is an expert in memory tuning and qualification for its GPU department, someone who is ready to work on GDDR7 memory for Intel's graphics division... well then.

NVIDIA launched its new GeForce RTX 50 series GPUs earlier this year, with new GDDR7 memory inside, while AMD launched its new Radeon RX 9070 series GPUs with GDDR6... but it looks like Intel is scoping out GDDR7 memory chips for its future Arc GPUs according to a new job listing spotted by @Haze2K1 on X.

The new job posting is looking for someone in Toronto, Canada with the job description reading: "The memory sub-system validation engineer is responsible to drive and deliver pre- and post-silicon validation and characterization of GDDR6/7 on Intel's ARC products".

NVIDIA could soon see its customized GeForce RTX 5090D graphics card banned in China, removed from shelves that would see gamers in the country shedding tears.

The Trump administration just updated its US export restrictions, immediately banning NVIDIA's tweaked H20 AI GPU from entering China, causing NVIDIA to write off $5.5 billion of revenue for Q1 2025. But now, according to a new post on the Chiphell forums, we're hearing that NVIDIA has suspended the sales of its highest-end RTX 509D in China.

Not only that, but NVIDIA has reportedly talked to AIBs to halt any and all sales of their custom GeForce RTX 5090D graphics cards in preparation for the new bans. The post continues, saying that this is a "precautionary" move by NVIDIA, as the situation between the United States and China is (very much) on-going.

Intel has acknowledged the performance loss between specific GPU and CPU pairings that was discovered several months ago, with the company now confirming an investigation has begun.

The problem was highlighted by Hardware Unboxed back in January, when a video was published honing in on performance loss when Intel's Arc B580 was paired with older processors such as the Ryzen 200, and Intel's own Core 9th-gen processors. Since then, Intel has remained quiet on the issue, but a new statement from the company, in the form of a forum post, has acknowledged the problem and confirmed that an investigation into the matter is underway.

Intel wrote, "We are aware of reports of performance sensitivity in some games when paired with older generation processors," and that it's increasing its "platform coverage to include more configurations in our validation process." This statement seemingly reveals Intel was aware of the issue prior to this forum post, but chose not to acknowledge it publicly. As for a remedy for the situation, Intel will have to roll out driver updates to the hardware in the form of patches that have an unknown release date or even release window.

Alongside the launch of the GeForce RTX 5060 Ti, check out our full review of the MSI GeForce RTX 5060 Ti Ventus 2X 16GB here, NVIDIA has also released the latest GeForce Game Ready 576.02 WHQL driver. In addition to supporting the new RTX 5060 Ti, this driver update introduces a very long list of game and general bug fixes that will interest many PC gamers.

Since the debut of the GeForce RTX 50 Series earlier this year, with the flagship GeForce RTX 5090, there have been increasing reports of users encountering 'black screen' issues, forcing a complete system reboot alongside stuttering and performance issues when connecting up more than one display to their GPU. Not limited to the new generation of RTX Blackwell cards, the problems have also been reported by those with GeForce RTX 40 and RTX 30 Series GPUs.

It's gotten to the point where game developers have recommended players roll back to drivers from as far back as December 2024. The good news is that GeForce Game Ready 576.02 features the biggest list of fixes we've seen in an NVIDIA driver update in a long time, covering things like game crashes, stuttering, black screen issues, multi-display problems, and more.

Zotac has revealed an RTX 5060 graphics card that'll be perfect for small PCs, and as you might imagine it's a very compact effort.

So much so that it features just a single fan, hence the name which is the Zotac Gaming RTX 5060 Solo - the 'Solo' referring to the presence of just one fan.

The board is very short as a result, but not short on performance, Zotac promises in a press release (via VideoCardz), and it should fit nicely in a small-form factor PC case (with an ITX motherboard).

Alongside the GeForce RTX 5060 Ti, NVIDIA has announced that the mainstream GeForce RTX 5060 will launch in May, starting from $299. With an MSRP of $299, NVIDIA is keeping the RTX 5070 price in line with the RTX 4060's launch price while also being cheaper than the RTX 2060 and RTX 3060 when they launched. In the current economic climate, this is impressive stuff; however, it remains to be seen how this translates to on-shelf pricing.

As for performance, the GeForce RTX 5060 is set to be a 1080p gaming powerhouse, delivering 100+ FPS in some of the most demanding titles. With max settings using DLSS Quality mode with Multi Frame Generation enabled, this translates to 114 FPS in Alan Wake 2, 130 FPS in Black Myth: Wukong, and even 130 FPS in the impressive Half-Life 2 RTX Remix demo.

NVIDIA notes that with Multi Frame Generation, the GeForce RTX 5060 can double the performance of the GeForce RTX 4060 while also lowering the latency. The following chart showcases the official GeForce RTX 5060 benchmark performance, with a couple of titles showcasing what to expect with Multi Frame Generation disabled.

NVIDIA has officially announced the GeForce RTX 5060 Series, with the GeForce RTX 5060 Ti launching on April 16. As the rumors have hinted, the GeForce RTX 5060 Ti will arrive in two flavors - 8GB or 16GB of GDDR7 on a 128-bit bus that delivers roughly double the memory bandwidth of the RTX 4060 Ti. With prices starting from $379 and $429, these mainstream RTX Blackwell GPUs benefit from the new architecture and the arrival of DLSS 4.

NVIDIA has released a single performance chart for the GeForce RTX 5060 Ti, which you can see above. The chart compares its performance to the GeForce RTX 4060 Ti and RTX 3060 Ti for 1440p gaming. Yes, the chart includes Multi Frame Generation, exclusive to the RTX 50 Series. Delivering 2X the frame rate with lower latency, it can deliver massive improvements to performance in games like Alan Wake 2, Black Myth: Wukong, and Cyberpunk 2077.

The chart includes two games running with like-for-like settings - Delta Force and A Plague Tale: Requiem. It shows a 20% or so performance improvement for the RTX 5060 Ti compared to the RTX 4060 Ti, which is good to see and a similar performance jump when looking at the RTX 4070 compared to the new RTX 5070.

About a week ago, we reported on a new rumor that AMD was planning to launch a Radeon RX 9070 GRE GPU in the Asian market. The GPU would feature 12GB of GDDR6 memory on a 192-bit bus with a cutdown version of the Navi 48 die found in the Radeon RX 9070 and RX 9070 XT. According to the initial report, AMD was already working with its partners to create cards for a third RDNA 4 GPU using the flagship Navi 48 die.

Well, thanks to GPU-Z's latest update, v2.65.0, which dropped over the weekend, the Radeon RX 9070 GRE is real and will soon hit shelves across Asia. GPU-Z GPU support typically arrives before a product launch or in line with a release. As the latest update also adds support for NVIDIA's GeForce RTX 5060 Ti and other RTX Blackwell laptop GPUs out shortly, it seems that the Radeon RX 9070 GRE will arrive before the Radeon RX 9060 XT.

This brings up the question of why. With the Radeon RX 9060 XT set to compete with the GeForce RTX 5060 Ti, recent specs and leaks suggest that the mainstream RDNA 4 GPU will feature half the GPU hardware as the Radeon RX 9070 XT. AMD may feel that the performance might not be there, and the Radeon RX 9070 GRE will be a better GeForce RTX 5060 Ti competitor.

Some new specs and hardware details surrounding AMD's Radeon RX 9060 XT have surfaced. This will be the company's answer to NVIDIA's upcoming GeForce RTX 5060 Ti. Like the RTX 5060 Ti, the more mainstream-focused RDNA 4 GPU is set to arrive in 8GB or 16GB flavors - however, with slower GDDR6 memory.

The latest piece of information arrives via Videocardz. Citing information from board partners, it outlines a few interesting things about the new Radeon RX 9060 XT. The reference spec includes a Game Clock of 2620 MHz and a Boost Clock of 3230 MHz. According to the report, OC models will push this to 3.3 GHz out of the box with a PSU requirement of 550W - which means some models might feature two 8-pin power connectors.

The updated specs also note that the GPU inside the Radeon RX 9060 XT, Navi 44, is essentially half the size of the GPU inside the Radeon RX 9070 XT - with 2048 Stream Processors compared to 4096 Stream Processors.

NVIDIA's RTX 5060 Ti graphics cards theoretically launch tomorrow (and go on sale the following day), and there's further buzz that they'll be priced more competitively than expected - but there are some big old catches here.

Moore's Law is Dead (MLID) brings us a new pricing and supply leak in his latest video, and it's very much good (well, kind of) and bad news.

The positive facet of this leak is that those rumors of a $379 MSRP for the RTX 5060 Ti 8GB and $429 for the 16GB flavor are seemingly true, at least according to one source at a major online retailer that MLID spoke to.

With increased power usage and a shift to a new power connector with the GeForce RTX 40 Series, one that has since been revised to 12V-2x6 for the GeForce RTX 50 Series, the arrival of RTX Blackwell has not been free from reports of melted power cables. The GeForce RTX 5090 is the most power-hungry gaming GPU to date, with a TGP of 600W traveling through a small cable.

Interestingly, the latest case of a melted or damaged power cable isn't related to the newest flagship or even the GeForce RTX 5080 but the 250W-rated GeForce RTX 5070. Based on various investigations and reports, this case looks different because one of the cables on the 12V-2x6 cable from an ATC 3.1 Seasonic 750W PSU has melted, not the connector.

Images and video of the aftermath, which includes a ZOTAC GAMING graphics card, were posted to X by @ere9w - where, upon closer inspection, the issue might be related to the connection on the GPU side.