We know it is coming, but we still do not know when. Intel has already confirmed that XeSS 3 Multi-Frame Generation will be released this month, and we already know which architectures will be supported. What we do not know yet is how well it performs on newer and older architectures. Thankfully, two reviewers have now shared new testing and impressions.
As we reported , a user
AMD has added a new graphics target, GFX1170, to the AMDGPU back end in the LLVM codebase. Phoronix first reported the change and noted the target is also labeled “ RDNA 4m .”
This ID was not know before, so it would fit between existing RDNA3 and RDNA4 families. Especially surprising given that AMD recently started enablement for GFX1251 (RDNA4 Refresh?)
Following tests on Windows, clearly showing Intel B-Series performance advantage over AMD iGPU, Phoronix tests also show the same trend in Linux gaming.
While it is not clear if we are going to see a spawn of new Intel-based gaming handhelds powered by SteamOS, those who prefer Linux experience and low-power graphics, should also know that Arc B-Series from Panther Lake is already showing pretty good results.
A
Digital Foundry found an unusual behavior in NVIDIA’s DLSS 4.5 Super Resolution. In testing, DLSS 4.5 presets M and L produced cleaner ray-traced reflections when the games’ own denoisers were disabled. The pixel hunting team mentions examples like Crysis 3 and Silent Hill 2, where disabling the built-in denoiser reduced “boiling” and blur in reflections.
Ray-traced effects often start from sparse samples,
MANGMI has announced the Pocket Max, its second Android handheld gaming device. The company is positioning it around a larger high refresh rate OLED display and Android plus streaming use cases.
MANGMI says the Pocket Max will start selling today through its official online store. The Super Early Bird price is $199.99 and the company says it is available from February 5 to February 12. MANGMI also lists an Early Bird
In a new video, Roman “der8auer” Hartung, YouTuber and CEO of Thermal Grizzly, said copper and tin pricing is starting to affect PC hardware supply and manufacturing costs. He described longer lead times and higher quotes for raw material used in CNC-machined cooling parts. More importantly, he explains how much more expensive it is to purchase copper for cooling companies.
Roman said an October order for 12 mm copper
A Nikkei Asia report says HP, Dell, Acer, and ASUS are considering memory chips from ChangXin Memory Technologies (CXMT) for the first time. The report ties that to a global memory shortage that is pushing up costs and complicating launch plans across the industry.
The report says HP has started qualifying CXMT parts as an extra supply option. It adds that HP could source CXMT for non-US markets if DRAM stays tight and pricing remains elevated through
HardwareLuxx reports that NVIDIA has now tied Dynamic Multi Frame Generation and Multi Frame Generation 6x to an April 2026 release window. The timing was shared during a company visit in Munich, according to Andreas Schilling.
At CES 2026, NVIDIA announced Dynamic Multi Frame Generation as a driver-controlled mode that aims for a user’s target refresh rate, instead of holding a fixed MFG multiplier. In the example given,
Please note that this post is tagged as a rumor .
NVIDIA’s reported GeForce RTX 50 SUPER refresh delay is resurfacing today with a new detail about what comes next. A report from The Information says NVIDIA managers decided in December to hold the refresh and prioritize AI chips due to tight memory supply.
While the report mostly repeats earlier reporting about the RTX 50 SUPER refresh delay , it does add one new point: it claims NVIDIA’s next
To be precise, it is not entirely correct to say GeForce NOW is six years old. The service is older than that. Readers who have followed VideoCardz from the start may recall that NVIDIA introduced GRID GPUs designed for cloud game streaming. This was part of a broader push for the SHIELD lineup, including set-top boxes, a controller, and the tablet. That dates back to 2013.
NVIDIA later rebranded the service as GeForce NOW