Aptly dubbed FSR 4, AMD says this delivers quality improvements to the technology across the board, using the benefits of machine learning much like Nvidia's offering. To achieve this, AMD has used custom ML models trained on AMD EPYC and Instinct accelerators to upgrade the AI upscaling algorithm of FSR 4.

Thanks to this, AMD is touting major image quality enhancements in FSR 4 compared to previous versions of the technology. Per the company's comparisons, it seems like FSR 4 can deliver better image fidelity in Performance Mode even over FSR 3.1's Balanced Mode at 4K.

On top of the frame rate gains seen just by using FSR 4, when AMD's own frame generation tech is applied to games, performance uplifts of up to 3.7 times could be seen as well. Check out the numbers that AMD showed off with several games below. To keep input lag at a minimum, AMD is once again using its Anti-Lag feature.

However, it doesn't look like previous generation AMD graphics cards will get this version of FSR, with the company saying FSR 4 is developed for the RDNA 4 Radeon RX 9070 series.

At launch, AMD expects to have over 30 games supporting the FSR 4 suite for RX 9070 series owners, while more than 75 titles are slated to add the technology later this year. It looks like the company is expecting fast adoption thanks to an upgradable FSR 3.1 API being offered to developers.

You can also find the entire coverage of our RDNA 4 (RX 9070 series) at this link.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

AMD says that it has added new dedicated math pipelines for ML acceleration that are primed for better performance with narrower data types. The new GPUs also gain support for FP8 (8-bit floating point) including BF8 (Brain Float 8)/FP8 E4M3 and HF8 (half precision float)/FP8 E5M2 for improved inferencing.

To leverage these improvements, the WMMA instructions have been enhanced on the 9070 series cards. In case you don't remember, AMD had added support for Wave Matrix Multiply Accumulate (WaveMMA or WMMA) last time with RDNA 3.

AMD is promising 2x the FP16 throughput on RDNA 4 vs RDNA 3 (RX 7000), and 4x in the case of INT8; and double those with Sparsity.

For Stable Diffusion AI image generation, the new RDNA 4 graphics are bringing AMD-optimized models with claims of much better performance thanks to corresponding memory usage improvements. The company also expects support for FLUX AI and LLMs in the future.

In real applications like Adobe Lightroom, Da Vinci Resolve, AMD is promising up to 34% better performance on 9070 XT vs 7900 GRE. Meanwhile, in the case of generative AI, the company expects the 9070 XT to be up to 70% faster than the 7900 GRE. These were measured on Windows 11.

They say with great power comes great responsibility, and AMD seems to be doing just that. Using all this AI firepower in its arsenal, the company is trying to further improve the stability of its display drivers with the new AMD AI Inspector feature.

The company is also finally bringing ML-based upscaling with FSR 4 and other frame-generation technologies

You can find the entire coverage of our RDNA 4 (RX 9070 series) at this link.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

Typically, gamers and creators have preferred Nvidia's NVENC encoder as a result of this since it requires AMD's AMF encoder to have higher bitrates in order to match the fidelity. This ultimately means that it gets resource-heavy to do the encoding work on a Radeon GPU.

With the latest RNDA 4-based RX 9070 series graphics cards, AMD is promising some major enhancements in this area on its VCN (Video Core Next). For example, compared to the 7900 XTX, Team Red is showing off much better quality while encoding H.264 (AVC) at 1080p using 6 Mbps bitrate. Speaking of H.264, the new media engine also improves low-latency H.264 encoding at lower bitrates.

In the case of HEVC (H.266), AMD claims an 11% encode quality improvement. AV1 is also seeing encoding efficiency improvements thanks to B Frames. Up to 8K 75 fps encoding is supported for HEVC and AV1, says AMD.

Aside from encoding improvements, decoding is also set to see some big gains. The company says the new VCN version brings low-power video playback with "more than 50% performance uplift for AV1 and VP9". This can be pretty significant since a lot of major web content is on VP9 or AV1. Not much is mentioned about H.264 and H.265 though. Perhaps the company feels there isn't much room for optimization for these codecs.

You can find the entire coverage of our RDNA 4 (RX 9070 series) at this link.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

Aside from the broad overview, the tech giant has also shared detailed performance numbers against the RX 7900 GRE. For those not familiar, the RX 7900 GRE sits between the 7800 XT and 7900 XT, and compared to Nvidia GPUs, the 7900 GRE is equivalent to the 4070 SUPER and 4070 Ti in raster, and about as good as a 4060 Ti in ray tracing.

The data consists of both 4K and 1440p resolutions using ultra (highest) graphics settings. AMD claims the 9070 XT is 42% faster than the 7900 GRE at the higher resolution and 38% faster than the GRE at the lower resolution.

At 4K (2160p), the new 9070 XT is said to be up to 28% faster at rasterization than the 7900 GRE with titles like Cyberpunk 2077 and Assassin's Creed Mirage showing the biggest gains. Speaking of Cyberpunk 2077, the 9070 XT should also see a huge improvement at ray tracing here with a 33% boost. Overall, AMD expects up to 34% better ray tracing.

Moving on to 1440p, the status quo alters slightly with more CPU-heavy titles like Assassin's Creed Mirage taking a back seat to games like Dragon Age: Veilguard, God of War: Ragnarok, and Black Myth: Wukong. Things don't change much on the ray tracing side though.

Up next, AMD has also published numbers for the RX 9070 which has eight fewer CUs and much lower clocks. Similar to the XT, the company has compared the RX 9070 to the next best in line, the 6800 XT and the Nvidia RTX 3080, and promises to deliver 38% and 26% better performance at 4K max settings respectively across 30+ games.

Detailed performance numbers of the RX 9070 have been shared again vs the 7900 GRE. AMD says that the 9070 is 21% and 20% faster than the GRE at 4K and 1440p respectively in overall performance across 30+ games.

Thus in a nutshell AMD expects some massive improvement per CU with RDNA 4. To learn more about how this level of performance was achieved, head over to this dedicated piece where the company has explained the underlying architectural enhancements.

As far as pricing goes, the RX 9070 XT is priced at $599 and the non-XT comes in at $549 SEP (Suggested Etail price). They will be available starting March 6th 2025.

You can also find the entire coverage of our RDNA 4 (RX 9070 series) at this link.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

The Radeon team says that the performance per Compute Unit (CU) per clock on RDNA 4 sees an uplift of nearly two times in rasterization compared to RDNA 2 (RX 6000 series) and by nearly 40% vs RDNA 3 (RX 7000 series).

Ray tracing sees an even bigger improvement as the 3rd Gen RDNA 4 ray accelerators claim to be about 2.4x or 140% faster than RDNA 2 and over 70% faster than RDNA 3. Thus, AMD is showing some impressive numbers here in terms of IPC (instructions per cycle).

However, AMD understands that raster and ray tracing are not everything in gaming these days as AI-based upscaling technology is also fairly common now. As such, Team Red's new GPUs are said to see the biggest gain in AI and ML processing. RDNA 4 promises up to four times faster FP16 dense matrix output vs RDNA 2, and against RDNA 3, it promises two times better performance.

AMD is moving back to a monolithic design with RDNA 4 as the new 9070 series cards are based on the TSMC 4nm process. There are in total 53.9 billion transistors across a die size of 356.5 sq.mm. area.

AMD has shared a high-level logical layout of the RX 9070 XT die which has 64 CUs in total arranged across four shader engines, where each shader engine comprises eight Workgroup processors. If you recall, each of AMD's RDNA Workgroup processors or WGP houses a couple of compute units.

With its new 3rd Gen Ray (RT) accelerators, RDNA 4 brings improvements to both ray traversal as well as shading performance.

The RX 9070 series is said to pack double the ray intersection rate. For those unfamiliar, the intersection rate is the process of determining how fast rays meeting a surface or object can be calculated on a GPU.

The new architecture also features significant improvements to BVH (Bounding Volume Hierarchy) primitive node compression to reduce memory bandwidth requirements. Along with that, RDNA 4 moves from a BVH4 to BVH8 implementation and this has been done to complement the doubling intersection rate. If you are wondering, BVH is a technique used to accelerate ray tracing performance.

And it does not stop there as false-positive intersections are also culled thanks to Oriented Bounded Boxes (OBB). AMD says that this new technology helps to boost ray traversal performance by around 10%, though it also depends on the source geometry. Overall, AMD says ray traversal IPC is twice as good on RDNA 4 compared to RDNA 3.

Following ray traversal, AMD has also highlighted the shading enhancements on its new GPUs. First up, the latency of memory requests is significantly reduced with a new Out of Order Memory returns technology. The technique makes out-of-order shader executions possible even in the case of cache misses and consequent wave request delays.

The new design also brings dynamic vector general purpose register (VGPR) management and allocation thus improving ray occupancy by lowering the number of idle VGPRs. This should immensely help ray shading since it typically requires more registers than ray traversal.

For some reason, AMD has not provided any details about the rasterization advancements as the focus this time is clearly on ray tracing and AI. In general, though, it does note that its RDNA 4 CUs feature "enhanced memory subsystem and improved scalar units" as well as "much higher clock speeds."

You can find the entire coverage of our RDNA 4 (RX 9070 series) at this link.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

The studio calls "Horizon Realms" one of the biggest updates for the game to date. New accolades, new vehicles, a new Stadium Track, Skill Mode, a fresh car pack, and more:

This April, the Horizon Festival evolves with Horizon Realms. This brand-new feature will let players dig into Forza Horizon 5’s history of Evolving World updates. Explore 12 fantastic Evolving World locations: 11 of the best previously limited-time locations and the premiering Stadium Track.

 

Collect new Horizon Realms-exclusive accolades to unlock four new-to-Horizon reward cars.  Revisit your favorite environments, or discover them for the first time!

Horizon Realms will bring back some of the user-favorite locations from the previous updates. Look out for the following 11 spots on the map:

• Stunt Park 
• Ice Rink 
• Oval Track 
• Summer Party 
• Neon Airstrip
• Día de Muertos 
• Winter Wonderland 
• Lunar Drift Arena 
• Retrowave Highway 
• Cars and Coffee 
• Stadium Maze

Each location will have five additional accolades to unlock (various tasks to complete in-game), which you will be able to track on the new Accolade tab. Also, the game will let you earn some of the accolades you might have previously missed.

Another addition is the brand-new Stadium Track and the Horizon Stadium Circuit.

As for the new cars, everyone will get a chance to earn the following new-to-the-game vehicles:

• 2024 Lamborghini Revuelto
• 2023 Lamborghini Huracan Sterrato
• 2018 Lotus Exige Cup 430
• 2022 Hennessey Mammoth 6x6

Also, a new car pack will be available for purchase on April 25. It will offer gamers four retro Nissan vehicles: the 1989 Nissan Pao, 1991 Nissan Figaro, 1989 Nissan Scargo, and 1987 Nissan Be-1. The pack will cost $4.99 and be available for purchase on April 25.

Finally, Forza Horizon 5 now has a release date on PlayStation 5. You can preorder the game, which will be available on April 29. It will offer a 60 FPS Performance mode and a 30 FPS Quality mode. Owners of the PlayStation 5 Pro will benefit from better visual fidelity and ray-traced car reflections in Quality mode.

Those preordering Premium Edition will get to play the game four days earlier, on April 25. Every preorder, regardless of the edition, gets you the 2021 Mercedes-AMG ONE, 5,000 Forzathon Points, and 5 Backstage Passes for purchasing hard-to-find cars.

You can read more about Horizon Realms in a post on the official Forza website.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

The integrated heat spreaders put on CPUs at the factory are not the most thermally efficient material you could have on there, but what are you going to do—rip it off at the risk of killing your $500 chip with your clumsy hands?

Yes, that is precisely what enthusiastic overclockers have been doing for years, delidding, or decapping (though the latter term is used less often in overclocking circles), chips through various DIY techniques, allowing them to replace AMD and Intel's common denominator shells with liquid metal or other advanced thermal interface materials.

As you might imagine, it can be nerve-wracking, and things can go wrong in just one second or one degree Celsius. In one overclocking forum thread, a seasoned expert noted that Intel's Core Ultra 200S spreader (IHS) needs to be heated above 165° C for the indium (transfer material) to loosen. But then the glue holding the IHS is also loose at this temperature, and there is only 1.5–2 millimeters of space between IHS and surface-mounted components, so it's easy for that metal IHS to slide off and take out a vital component with it. It's quite the Saturday afternoon hobby.

That is the typical overclocking bargain: You assume the risk, you void your warranty, but you remove one more barrier to peak performance. Now, though, Thermal Grizzly, led by that same previously mentioned expert, Roman "der8auer" Hartung, has a new bargain to present. His firm is delidding AMD's Ryzen 9800X3D CPUs with its own ovens and specialty tools, then selling them with two-year warranties that cover manufacturer's defects and "normal overclocking damage," but not mechanical damage.

der8auer explains his rather unique delidded CPU offering and shows you how it is made.

In a video explaining this business for which almost no bank would ever provide capital, Hartung/der8auer makes it clear that "AMD has no clue what we're doing here" and that AMD will not help his business with anything. Thermal Grizzly is buying the CPUs from German store Mind Factory, spending the labor time to delid, test, and verify each chip, then selling them at a markup. That markup will mostly cover returns, as Thermal Grizzly has to eat the more than 500 euro cost of any chips that come back under warranty.

The package you get from Thermal Grizzly when ordering a delidded AMD Ryzen 9800X3D, minus the USB drive.

Credit: Thermal Grizzly

You can also see the process by which batches of chips are put into a precision oven and baked, first at 170° C to loosen the indium and glue, then again upside-down at 160° C to let the IHS fall off. "It smells like CPU barbecue in here a little bit," Hartung says, while what he describes as "$10,000 worth of CPUs is going to be melted." After that, the chips are cleaned up, slotted into a testing system, run through a Cinebench R23 test, and monitored for their temperatures. Each customer receives the chip, the IHS, a USB thumb drive with a detailed shot of their chip and its IHS lid, and a verification card stating the temperatures the chip reached during its testing.

Thermal Grizzly's delidded Ryzen 7 9800X3D chips are currently going for $712.95 in the US (VAT and shipping included) or about $230 over the official $479 suggested retail price. The chips are sold out as of this post's publication, but more will likely come available as they bake them. Hartung states in the announcement video that his firm may expand to other chips in the future, presumably after he finds out how many people like a slightly less risky kind of tinkering.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

What's notable about AWS' Ocelot, quantum computing chip is that it reduces the cost of implementing error correction by up to 90%. AWS has published its findings in a paper on Nature. By reducing costs, and publishing findings, this development is a step towards democratizing access to quantum computing.

“With the recent advancements in quantum research, it is no longer a matter of if, but when practical, fault-tolerant quantum computers will be available for real-world applications. Ocelot is an important step on that journey,” said Oskar Painter, AWS director of Quantum Hardware. “In the future, quantum chips built according to the Ocelot architecture could cost as little as one-fifth of current approaches, due to the drastically reduced number of resources required for error correction. Concretely, we believe this will accelerate our timeline to a practical quantum computer by up to five years.”

According to AWS, Ocelot contains 14 core components. There are five data qubits (cat qubits), five buffer circuits to stabilize the data qubits, and four more qubits for detecting errors on the data qubits. For anyone who wants to get further into the nitty-gritty details, Amazon has published a technical article about Ocelot.

Quantum computing is a pretty confusing topic, but within the next decade, we are going to be hearing much more about quantum computing breakthroughs, so it's a good idea to get a basic grasp on the topic. To this end, Neowin published an editorial explaining the basics of quantum computers and what they'll unlock.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

The Nvidia hardware-exclusive feature set brings Multi Frame Generation, Frame Generation, Ray Reconstruction, Super Resolution, DLAA, Reflex Low Latency, and NVIDIA Image Scaling as a part of the DLSS suite.

Keep in mind that despite supporting graphics cards going back to the RTX 20 series, not all of DLSS 4's features are available to those generations.

Here are the DLSS 4 technologies that are using neural rendering techniques, and what they do exactly:

• DLSS Multi Frame Generation boosts frame rates by using AI to generate up to three frames per rendered frame, powered by GeForce RTX 50 Series and fifth-generation Tensor Cores.
• DLSS Frame Generation boosts performance by using AI to generate frames while maintaining great responsiveness with NVIDIA Reflex.
• DLSS Ray Reconstruction enhances image quality by using AI to generate additional pixels for intensive ray-traced scenes. DLSS replaces hand-tuned denoisers with an NVIDIA supercomputer-trained AI network that generates higher-quality pixels between sampled rays.
• DLSS Super Resolution boosts performance by using AI to output higher-resolution frames from a lower-resolution input. DLSS samples multiple lower-resolution images and uses motion data and feedback from prior frames to construct high-quality images..
• Deep Learning Anti-Aliasing provides higher image quality with an AI-based anti-aliasing technique. DLAA uses the same Super Resolution technology developed for DLSS, constructing a more realistic, high-quality image at native resolution.

So far, games like Cyberpunk 2077, Indiana Jones and the Great Circle, Alan Wake 2, Marvel Rivals, and Star Wars Outlaws have shipped updates with support for DLSS 4. With an official plugin now available for Unreal Engine 5 though, expect much more upcoming games to include the upscaling and frame generation features for gamers.

The DLSS 4 plugin for Unreal Engine 5 supports versions 5.5, 5.4, 5.3, and 5.2. Find the download and resources to learn about DLSS and its implementations here, Nvidia's dedicated developer portal.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

This week, Microsoft unveiled a new brain for quantum computers called Majorana 1. The chip, which fits in the palm of your hand, leverages what are known as topological qubits – it sounds very complicated, many people have pointed this out, but the main benefit of this type of qubit is its innate error resistance.

The Redmond giant explained that the chip currently holds 8 qubits but the design makes it possible to scale this to 1 million qubits in the coming years. With this breakthrough, practical quantum computers are no longer decades away, but years.

In this editorial, I will take a look at the benefits of quantum computing, the history of quantum computing, explain Majorana 1 and where it fits into Microsoft’s quantum computing roadmap, and speculate on how far away quantum computing is, according to industry leaders.

The benefits of quantum computers

We’ve all been hearing about quantum computers for a long time now, but most people, I would wager, find the topic quite dense and impenetrable. The key point around quantum computing, is that when these computers are built, and capable of practical workloads, they’ll be able to solve some problems much more quickly than classical computers.

In classical computers, the basic units of data are called bits. They are binary, meaning at any one time they can have a value of 0 or 1. Classical computers run calculations and process data based on these bits and use logical operations to manipulate them. These 0s and 1s are represented by the switch position of transistors on microchips – for reference, microchips in modern smartphones have over 10 billion transistors.

Leveraging transistors, we get logic gates, the building blocks of classical computing. They perform operations on bits such as AND, OR, and NOT to create more complex functions. The final basic point about classical computers to note here is that they perform instructions in order, one at a time, which limits their speed and efficiency in some tasks.

With quantum computers, the basic unit is called a qubit, they have different properties to bits in classical computing which can make quantum computers much more efficient in certain areas. Unlike the bit, the qubit can have a value of 0, 1, or both at the same time – this is a principle from quantum mechanics called superposition and allows for simultaneous processing.

Another important property of quantum computers is entanglement. This refers to the ability of qubits to become entangled with other qubits so that they affect each other's state, this connection between them allows for faster information processing and complex problem-solving.

The final fundamental concept in quantum computing that I want to mention is called interference. After a quantum computer has used superposition and entanglement to create a vast number of possible states, it uses a principle called interference to amplify and suppress states. This further boosts the efficiency of computations.

To unlock these unique abilities of qubits, quantum computers use quantum gates, in addition to logic gates found in classical computing. Quantum computers enable quantum algorithms such as Shor’s Algorithm and Grover’s Algorithm, which efficiently factors large numbers and speeds up unstructured search problems, respectively.

Put simply, quantum computers can solve some problems much faster, are capable of parallelism, and help humanity to tackle problems that are currently unsolvable with classical computers.

When quantum computers for practical tasks emerge, it is expected that they’ll have a significant impact on several fields including cryptography, drug discovery and materials science, artificial intelligence and machine learning, financial modeling, as well as climate modeling and weather forecasting.

In the field of cryptography, experts are concerned that quantum computers could possibly break today’s encryption methods. However, this technology could also enable new forms of cryptography that’d be secure against quantum computers and classical computers.

Quantum computing’s ability to quicken drug discovery and materials science is one area where people would be impacted in their everyday lives for the better. Quantum computers can more accurately simulate molecular interactions, this could significantly speed up drug discovery processes and the development of new materials.

Quantum computers also hold potential in finance for creating better investment strategies and optimizing portfolios by analyzing difference scenarios and outcomes simultaneously.

Industries that stand to benefit from the development of practical quantum computing include healthcare, banking and finance, logistics and supply chain management, telecommunications, energy, AI, and manufacturing. The increased efficiency could also lead to cost reductions that could be passed onto consumers through the mechanism of competition between market participants.

A brief history of quantum computing

The idea of a quantum computer was proposed in the 1980s with no real consensus on who came up with it first. American physicist Paul Benioff described a quantum mechanical model of a computer that employed quantum mechanics in a 1980 paper, and Richard Feynman, the well-known American physicist, proposed the idea of using a quantum computer to simulate the behavior of physical systems. Feynman’s paper was released in 1982, he recognized that classical computers would struggle to simulate quantum systems and proposed a quantum computer to solve this.

Another significant contributor to the field of quantum computing is British physicist and computer scientist David Deutsch. He made some notable contributions including the concept of a quantum Turing machine, quantum parallelism, and quantum error correction.

In the 1990s, we got the creation of Shor’s Algorithm (1994) and Grover’s Algorithm (1996), which we mentioned earlier. Shor’s Algorithm allows for the factoring of large numbers while Grover’s Algorithm improves the efficiency of searching an unsorted database. This decade also saw the development of small-scale quantum computers from IBM and MIT, among others.

During the early 2000s, Microsoft began its research into topological quantum field theory, laying the foundations for its work with topological qubits, and ultimately, the development of the Majorana 1 chip.

In 2011, D-Wave, a Canadian company, made headlines with the D-Wave One, which claimed to be the world’s first commercially-available quantum computer. Fast-forward to the end of the decade and Google has built a quantum computer called Sycamore which achieved quantum supremacy in 2019.

Quantum supremacy is the idea that a quantum computer solves a problem that no classical computer can in a reasonable amount of time. Google achieved this with the 53-qubit Sycamore in 2019 by solving a problem in just 200 seconds what it’s believed would have taken a classical computer 10,000 years.

In the 2020s, efforts have continued towards practical quantum computers with firms like IBM, Google, and Microsoft leading the way. Each of these three companies has outlined their goals for the coming years, with developments expected to culminate around 2035.

With the 2010s seeing quantum computers becoming commercially accessible, it also meant that developers needed tools to create software. We’ve seen the launch of quantum software frameworks and quantum programming languages. Within Microsoft’s quantum computing ecosystem, there is the Q# language, which uses high-level syntax for writing quantum algorithms that run within Microsoft’s quantum platform.

IBM also has a quantum programming language called OpenQASM which can be used in addition to Python and Rust in the Qisket software development kit. Meanwhile, Google has developed the Cirq framework where developers can use Python to create quantum circuits. Many of these tools allow developers to run the programs in a simulator before running it on a cloud-based quantum computer.

Microsoft’s quantum computing roadmap and Majorana 1

On Wednesday, February 19, 2025, Microsoft unveiled its Majorana 1 quantum computer chip that, right now, holds 8 topological qubits, but, over time, can be scaled up to 1 million qubits. Chips that use topological qubits are different to what Microsoft’s rivals are building and have the benefit of being more error resistant than other types of qubits.

Error resistance is very important in quantum computing because they’re inherently prone to errors due to the fragile nature of quantum states. As Microsoft scales up the number of qubits on its chip, the error resistant properties of topological qubits will start to shine and reduce the obstacles the Redmond giant faces in the future.

Before discussing Microsoft’s roadmap, I want to quickly provide information about why the company called its chip Majorana 1, as some people online have questioned the choice of name.

Microsoft’s chip is named after the Italian physicist Ettore Majorana who came up with the concept of Majorana particles. These particles, or fermions, make up the building blocks of topological qubits. The error resistance in topological qubits comes from the fact that Majorana particles are their own antiparticles which allows them to encode data that’s more error resistant.

With the Majorana 1 chip, Microsoft has achieved the second stage of its six-stage roadmap towards practical quantum computers. The stages of the roadmap are as follows:

MILESTONE 01: Create & control Majoranas

 

For the first time in history, Microsoft engineered devices allow us to induce and control the topological phase of matter bookended by Majorana Zero Modes. This breakthrough enables the engineering of a new type of qubit.

 

MILESTONE 02: Hardware Protected Qubit

 

Our protected qubit, with built-in error protection, extends our first breakthrough by changing qubit technology from analog to digital control.

 

MILESTONE 03: High Quality Hardware Protected Qubits

 

To scale operations and reduce errors, digitally controlled hardware-protected qubits can be entangled and braided with a series of quality advances.

 

MILESTONE 04: Multi-qubit System

 

A variety of quantum algorithms can be executed when multiple qubits operate together as a programmable Quantum Processing Unit (QPU) in a full stack Quantum Machine.

 

MILESTONE 05: Resilient Quantum System

 

A Quantum Machine, when operating on true logical qubits, demonstrates higher quality operations than the underlying physical qubits. This breakthrough enables the first reliable quantum operations and opens the gates to quantum supercomputing.

 

MILESTONE 06: Quantum Supercomputer

 

The Quantum Supercomputer solves scientific or commercial problems faster than classical computers, starting at 1 million reliable rQOPS/sec with an error rate below 1 in a trillion, scaling to 100 million rQOPS/sec for advanced chemistry and materials science challenges.

To get from the first to the second stage of the roadmap, it took Microsoft 18 months. The company has stated that it’ll reach its goal in years rather than decades. This suggests that the latest date by which it will reach its goal will be around 2035, however, if we assume it takes 18 months to reach each goal, it could be completed by 2031.

The work Microsoft is doing with Majorana 1 has attracted the attention of the Defense Advanced Research Projects Agency (DARPA). Microsoft is one of just two companies in the final phase of DARPA’s Underexplored Systems for Utility-Scale Quantum Computing (US2QC) program. The program makes up part of the Quantum Benchmarking Initiative which aims to deliver “the industry’s first utility-scale fault-tolerant quantum computer”.

How far away are quantum computers?

Predicting the arrival of quantum computers is tricky, simply, we don’t know exactly when machines with a million stable qubits will arrive. Right now, there are machines in the low hundreds of qubits, but a million is still quite far off. Reducing the number of errors in quantum computers, and cooling, are also significant issues that need to be overcome.

Similarly to Microsoft, Google also has a six-stage timeline and so far, the first two milestones have been hit; the first was hit in 2019, and the second was in 2023. At the 2023 milestone, Google said its quantum computer has 100 physical qubits. The aim in milestone 3, 4, 5, and 6, respectively, is to have 1,000, 10,000, 100,000, and 1,000,000 physical qubits.

If Google reaches subsequent milestones at the same rate it took to get from the first to the second milestone, then milestone 3 will be reached in 2027, M4 in 2031, M5 in 2035, and M6 in 2039. That’s not too far off what Microsoft is saying about quantum computers being years away, not decades.

IBM states in its roadmap that it will have a quantum computer around 2033 capable of running 1,000s of logical qubits. Logical qubits are actually encoded across lots of physical qubits in an attempt to reduce the number of errors. Google is taking this approach too, but Microsoft’s topological qubits are designed to be inherently stable, which could give the Windows maker the edge in this race.

Before we get there, a number of hurdles need to be overcome, chiefly error rates and decoherence. There is also the issue of scalability; the quantum computers around today only have small numbers of qubits and as companies try to add more, it will add to the complexity of systems.

On the hardware side of things, for practical quantum computers to become a reality, there will need to be new materials developed such as new semiconductors, better insulators, and topological insulators. These will need to be stable, cheap, and reproducible.

Quantum computers are also expensive to run due to the cooling they require, limiting their deployment to governments and big tech companies. To make them more accessible, there will need to be a miniaturization of setups, and new ways unlocked for them to operate at higher temperatures so they don’t have to be so intensively cooled.

Finally, there is the issue of standardization. As you may have gathered from this editorial, the big tech firms are all approaching the quantum computing problem from multiple angles making their systems incompatible with each other. In the long term, there will have to be some sort of convergence around common standards.

Conclusion

In this editorial, I have provided an outline of the benefits of quantum computers, a brief history of quantum computers, explained why Microsoft’s Majorana 1 is notable, and had a look at how far away we are from practical quantum computers.

The impetus for this editorial was the unveiling of Majorana 1. I think it’s a very worthwhile development, given the fact that Microsoft has shown off a new type of qubit that attempts to solve error issues, while offering a seemingly viable path for scaling to 1 million qubits. The Redmond giant’s prediction that quantum computers are no longer decades away suggests that optimistic timelines could turn out to be correct, which is an exciting development.

I think the potential that quantum computers hold is massive, with the anticipated improvements in healthcare, materials, and climate modeling. When I first tried ChatGPT in 2022, it felt like magic. I think that quantum computing, while it probably won’t be directly in our hands at first, will also lead to similar magnitude developments in the mentioned fields and beyond.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

Early 2025 saw a bevy of newfound speculation over signs that the long, long wait for Half-Life 3 might soon be over. Now, data contained in some new Valve game updates suggests that the project known in Valve engine code as "HLX"—and widely assumed to be Half-Life 3—might be reaching the final stages of production.

In a new video, longtime Valve watcher Tyler McVicker goes into detail on a bevy of new variables and strings found after spending hours datamining the latest update to Dota 2 (the first update for that game since mid-December). The strings suggest a wave of behind-the-scenes Source engine changes dealing with the kind of "optimization and polish" that "happen(s) at the end of a game's production cycle," McVicker says. "This is getting to the point where it does feel as if Valve is nearing completion of the production of HLX."

Tyler McVicker goes over all the new datamined evidence that "HLX" development is wrapping up.

Those changes include a set of new code in a file called AI_baseNPC.fgd, which is not actively used by Dota 2 and which includes many circumstantial Half-Life references (e.g. "machinery," "alien blood"). The specific code in this latest update deals with letting the engine scale the level of an NPC's AI simulation based on its distance from the player, a refinement that McVicker says is "absolutely... optimization work" and an apparent sign that "Valve has hit the optimization and polish phase" on HLX.

McVicker also makes note of a February 25 update to Deadlock which includes a new variable suggestively named "hlx_fsr3_min_reactiveness." That's being taken as an apparent sign that AMD's AI-powered FidelityFX Super Resolution upscaling is in use in a current HLX development build, which would be another sign that HLX "legitimately is nearing the end of development," as McVicker puts it. "You don't use FSR until you're nearly done with a game. Those AI post-processing systems are not supposed to be used for development."

Don’t you dare give me hope

The newly discovered engine updates are a particularly good sign for the HLX project after credible reports that the game had gone into a friends and family playtesting phase at Valve near the end of 2024. While many Valve projects have seen major changes after that kind of testing, McVicker says that these new engine updates are "confirmation indeed that HLX survived the holiday break, does not seem to be going through any form of reboot, and [that] an announcement seems to be inching ever closer."

A December video from "Gabe Follower" suggesting that Half-Life 3 had entered a friend and family testing phase.

While McVicker doesn't go into detail on every single engine variable hinted at in the latest Valve game updates, he provides a summary version with informed speculation on what all those variables could mean for HLX. That includes signs of a "heavily evolved" physics engine from Half-Life 2 that can seemingly track object traits such as buoyancy, flammability, deformation, as well as fluid simulation and dynamic sound properties based on an object's status.

Most intriguingly, McVicker talks up signs that HLX may include a new NPC "mood system," letting these AI-controlled characters react "verbally and non-verbally" to what they can see, hear, or even smell in the immediate area. The code also hints at "dynamically adjustable nav meshes" that will let NPCs navigate easily around destructible environmental hazards, and at a system for in-game objects to be placed semi-randomly during multiple playthroughs, a la Left 4 Dead's director system.

We know it's easy to be skeptical after so many years of fruitless speculation regarding Half-Life 3. But when McVicker says that signs suggest that an actual Half-Life 3 announcement could come this year, let's just say we want to believe.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

A megathread on the ASRock subreddit reveals that a significant number of these processors appear either dead on arrival (DOA) or fail after only a few hours, days, or weeks of use. Furthermore, moderators have confirmed that over 40 Ryzen 7 9800X3D units have reportedly been affected, with a few non-X3D processors also falling victim. This raises the possibility that similar issues may extend to other Zen 5 CPUs like the upcoming 9950X3D, although this has yet to be verified.

The reliability issues are particularly troubling for gamers, who have eagerly awaited AMD’s X3D series amidst already tight supply. The emergence of these problems only adds to the frustration, as the performance potential of these CPUs is being overshadowed by inconsistent reliability.

In response to the mounting reports, ASRock has released a new BIOS update (version 3.20 Beta) for its AM5 series motherboards. This update is designed to address unexpected boot issues and error codes reported by users of AMD Ryzen 9000 series processors. Notably, the update’s accompanying blog post does not explicitly mention CPU failures, suggesting that the focus is on boot-related issues rather than the DOA incidents reported with the Ryzen 7 9800X3D.

In addition, a statement from ASRock Japan has hinted that the CPU failures might be linked to memory compatibility issues like those related to memory training perhaps, rather than a fundamental flaw in the processor architecture. However, Team Red has not yet provided an official comment on these reports, leaving the root cause somewhat uncertain.

Here's what ASRock Japan has said (Google-translated to English):

We will release an official release with more details later, but since there is a lot of misinformation being spread, here is a quick note! The bottom line is that it is not broken. It is just that some CPUs do not start due to memory compatibility issues. There are many reports of Ryzen 7 9800X3D suddenly breaking. At least 40 cases. Occurring mainly in ASRock motherboard environments

For the time being, users experiencing these issues are advised to update their motherboard BIOS as per the guidelines provided by their manufacturers. AMD has acknowledged the problem and is currently accepting RMA requests for affected units. As the situation develops, current and prospective buyers should continue to monitor updates from AMD and their motherboard vendors.

Stay tuned for further updates as more information becomes available.

Via: Wccftech

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

While Nvidia later confirmed that it had found a manufacturing issue affecting the RTX 5090/5090D and 5070 Ti models, not even the RTX 5080 was found to be safe. User reports from earlier this week found that this high-end can also be affected by the missing hardware problem and loss of raster performance.

Now, Nvidia has confirmed that the issue is indeed affecting some RTX 5080 GPUs.

In a statement to The Verge, Nvidia GeForce global PR director Ben Berraondo said, "Upon further investigation, we’ve identified that an early production build of GeForce RTX 5080 GPUs were also affected by the same issue. Affected consumers can contact the board manufacturer for a replacement."

Considering that this statement also arrived after reports by users and hardware publications and not preemptively by the company, The Verge had questioned if the missing ROPs issue can pop up in any other Nvidia GPUs. To this, Berraondo has added that "no other Nvidia GPUs have been affected" and that “the production anomaly has been corrected," specifically mentioning the upcoming RTX 5070.

According to Berraondo, Nvidia had not known about this hardware issue prior to the launch of the RTX 50 series.

Nvidia is asking users affected by the problem to contact their board manufacturers for a replacement. Considering the stock issues the RTX 50 series has been having, though, it's unclear how long the defective card owners will have to wait for a new GPU.

The easiest way to find out if your RTX 50 series card has missing ROPs is to use GPU-Z and look at the available ROPs section. This should read 176 on RTX 5090s, as listed in the specifications by Nvidia. On the RTX 5070 Ti, it should read 96 ROPs, while defective units will read 88 ROPs. Lastly, the RTX 5080 should ship with 112 ROPs, but defective units will read 104 instead.

The latest problem arrives to join several others, like melting connectors and black screen issues, being faced by RTX 50 series owners. The company also stopped supporting 32-bit PhysX with the latest generation cards, which had even prompted one user to purchase and pair a whole other RTX 3050 graphics card with an RTX 5090, Nvidia's most expensive and most powerful consumer offering, to just run the PhysX calculations on classic titles without losing massive performance.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

We’re on the cusp of the release of my most anticipated game of the year, Monster Hunter Wilds. For years, I’ve been waiting for this follow-up to one of my favorite games. To give some context, I poured over 300 hours into Monster Hunter World.

With reviews now live, Monster Hunter Wilds has garnered mostly positive reception. However, our own Wilds review here at Windows Central criticized the game’s lack of difficulty—something at odds with the premise of the franchise for many long-time fans. Having had the privilege of playing the game early, I wanted to share my own thoughts, particularly some counterpoints.

While I agree that the reduced difficulty might feel like a departure for veteran players, I see it as an overall positive. It makes the game more accessible to newcomers and casual fans, something that may lead to more mainstream success for the franchise than ever before. That said, I have my own major frustration with Wilds.

The multiplayer experience, which is my main way of playing this series, remains as convoluted and clunky as ever. In 2025, I have far less patience for these outdated systems than I did in 2018 when Monster Hunter World launched. As a result, I don’t see myself dedicating nearly as much time to Wilds, and that’s a real downer for a game I’ve been so excited about.

I had higher hopes for the co-op experience

Gamers Nexus (2.4M subscribers)

February 25, 2025

Video length: 25m 21s

00:00 - The Worst GPU Launch

04:37 - NVIDIA's Response

07:10 - How This Might Have Happened (Speculation)

12:31 - Rant: DISASTROUS LAUNCH

18:36 - Our Theory's Problems: RTX 5080 Isn't Immune

19:51 - Conclusion: AMD's Opportunity

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

The GPU-Z screen capture below, shared by Reddit user gingeraffe90, clearly specifies the lower-than-intended ROP count. Instead of the regular 112 ROPs, the missing 8 ROPs are similar to those found on GeForce RTX 5090 and RTX 5070 Ti. Gingeraffe90 posted the screenshot using the latest version of GPU-Z.

The loss of eight ROPs constitutes around 7% of the degradation in the GPU's raster hardware capability. This means that scenarios where GPU is heavily dependent on operations like rendering high-resolution scenes will be impacted by a similar margin. However, it is important to note that gaming performance depends on varied factors, and the overall performance impact may differ from game to game.

The graphics card in the above scenario is Nvidia's Founders Edition, and the poster has performed GPU-Z validation after reinstalling drivers.

For reference, ROP is a hardware component and the final step in the rendering process of graphics cards, and it's not the first time the missing ROP issue has been reported. Nvidia has confirmed that it is a "rare issue" that affects GeForce RTX 5090/5090D and 5070 Ti cards, and only 0.5% of those GPUs are influenced. However, the company hasn't explicitly mentioned that RTX 5080 is also prone to the same problem. Also, Nvidia never revealed the reason behind this nuisance.

Since the post has merely been online, it will take time for Nvidia to investigate the issue and furnish their comment. Therefore, it is too early to draw any conclusions.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

With the emergence of generative AI, AI-powered tools like Copilot, ChatGPT, and more are quickly gaining broad adoption.

As a result, there's been a dramatic shift, especially in how people interact with information on the Internet, thanks to the emergence of AI-powered tools like OpenAI's ChatGPT search.

OpenAI CEO Sam Altman recently revealed that he's using ChatGPT more than ever as it seemingly evolves into a so-called 'Everything App.'

"I don't do Google searches anymore," added Altman.

A new study highlighted by our sister site, Tom's Guide, echoes the same sentiments. According to a survey featuring 510 participants in the US and an additional 518 in the UK conducted by our parent company, Future Publishing, approximately 27% of the US participants are heavily inclined to use AI-powered search tools like ChatGPT as conventional tools like Google seemingly fade away into the horizon.

AI specialist Amanda Caswell attributes the dramatic shift to well-curated, in-depth, and contextual answers delivered with a more human-like tone in a straightforward set of apps that save time, are easy to use, and offer personalization for added accuracy and relevance.

Interestingly, a separate study by Gartner seemingly corroborates Future's findings, citing that AI-powered tools like ChatGPT search will take over 30% of online search queries by 2026, potentially denting Google's longstanding search dominance.

It'll be interesting to see how companies at the forefront of the dramatic shift to AI tools for online search queries handle abound issues, including the rampant erroneous AI-generated summaries.

Satya Nadella could make up for Microsoft's Google transgressions with Copilot AI

The Cupertino giant said it will work with partners to manufacture servers in Houston, Texas, later this year. These servers, previously manufactured outside the US, will power Apple Intelligence and act as the foundation of Private Cloud Compute.

A 250,000-square-foot server manufacturing facility is planned to open in 2026. The company is all in on Apple Intelligence and plans to expand its data center capacity in North Carolina, Iowa, Oregon, Arizona, and Nevada.

“We are bullish on the future of American innovation, and we’re proud to build on our long-standing U.S. investments with this $500 billion commitment to our country’s future,” Apple CEO Tim Cook said in a press release.

Apple will grow its R&D investment in the US to support fields like silicon engineering. It will double its US Advanced Manufacturing Fund by increasing it from $5 billion to $10 billion. The fund was created in 2017 to support high-skilled manufacturing jobs in the country.

Apple is making a multibillion-dollar commitment to produce advanced silicon in TSMC’s Fab 21 facility in Arizona as part of the manufacturing fund's expansion, where Apple is the largest customer. The company notes its suppliers manufacture silicon in 24 factories spread across 12 US states, and the manufacturing fund has supported projects in 13 states.

Apple said it supports 2.9 million jobs in the US through direct employment, developer jobs in the iOS app economy, and work with US-based vendors. It plans to hire 20,000 more in the next four years, mostly in R&D, silicon engineering, software development, and AI/machine learning.

Speaking of R&D, Apple has already doubled its spending in the last five years. The recently launched iPhone 16e packs the first cellular modem designed by the company, called the Apple C1. Apple claims that the C1 is the most power-efficient modem ever in an iPhone.

While the modem may have contributed to the extended battery life, it lacks advanced 5G technologies such as mmWave. However, it's the first among the lot and Apple is working on a long-term strategy to fit in-house modems inside iPhones.

A new Apple Manufacturing Academy is also being setup in Detroit, offering free in-person and online courses to teach vital skills like project management and manufacturing process optimization.

It's not all geeky stuff; Apple's $500 billion massive package will fund Apple TV+ productions in 20 states. Apple's in-house production studio and streaming service have been around since 2019, producing several award-winning titles.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

Table of contents:

1. Windows 10 and 11 news
2. Windows Insider Program
3. Updates are available
4. Reviews are in
5. Gaming news

Windows 11 and 10

Here we talk about everything happening around Microsoft's latest operating system in the Stable channel and preview builds: new features, removed features, controversies, bugs, interesting findings, and more. And of course, you may find a word or two about older but still supported versions.

If you have a Copilot+ PC, you can check out the updated version of AI Hub. It is a dedicated storage page that now promotes various AI-powered experiences on your device, such as Windows Studio Effects, Live Captions, Paint Cocreator, and more. It also highlights applications with AI experiences. The updated AI Hub is now available to customers in the United States and a couple more countries.

This week's Windows 10 and 11 section is pretty slim, so here is a useful guide from Paul Hil, who describes how to create a bootable Windows USB device on a Linux-powered system.

Windows Insider Program

Here is what Microsoft Released this week for testing in the Windows Insider Program:

In addition to three new builds, Microsoft announced a new update for the Snipping Tool app. Version 11.2501.7.0, which is now available in preview for Windows Insiders, introduced a new cropping tool, which is handy for screen recordings—no more jumping to other apps just to crop out unnecessary parts of screenshots.

As usual, enthusiasts took apart the latest Windows 11 preview builds and discovered some interesting tidbits about upcoming features. Microsoft is apparently working on a new file migration app to help you transfer files between PCs without using third-party software. No announcements have been made yet, but you can check out an early UI reconstruction here.

Updates are available

This section covers software, firmware, and other notable updates (released and coming soon) delivering new features, security fixes, improvements, patches, and more from Microsoft and third parties.

Remember that story about Microsoft wanting to keep your Microsoft Account always signed in? As it turned out, that is not happening, at least yet. The company removed the support document and clarified that it was published prematurely. The change is still coming; we just do not know when.

Microsoft Edge received several updates this week. For one, Microsoft revealed that version 132, which arrived in January, updated several parts of the user interface to WebUI 2.0. This change brought significant performance and UI responsiveness improvements, especially on lower-end machines. Future updates will introduce WebUI 2.0 to more parts of the browser.

Microsoft also released a new feature update on the Dev Channel. Version 135 is now available for testing, and it includes various fixes that could cause the browser to crash.

That is not all about Microsoft Edge. Users in the Stable Channel received a security update that patched several high-severity vulnerabilities. Finally, Microsoft announced the release of a revamped Edge Add-ons website where you can download extensions for Microsoft Edge.

Mozilla released a small bug-fixing update for Firefox 135 (released earlier this month). It fixes issues with non-working drop-down menus on certain pages, scrolling bugs, and more. By the way, Mozilla also announced that Firefox 115 ESR will remain supported on Windows 7, Windows 8, and Windows 8.1 for six more months.

Finalwire has some bad news for Windows 98 users: the latest AIDA64 release no longer works on Windows 95, 98, and Me. Version 7.60, which arrived this week, dropped support for truly ancient Windows releases while also introducing support for AMD's upcoming graphics cards and other changes.

Next, we have a new version of the Battery Flyout app, a handy tool for tracking and managing your devices' batteries. With the latest release, Battery Flyout can generate battery reports with detailed information about health, number of charges, and more.

Also, those who are discontent with Windows 11's File Explorer can check out File Pilot, a new third-party file manager with some seriously impressive performance. At just 1.8MB, File Pilot offers a massive list of features and crazy responsiveness. It is now available for free in beta, but the final release will be paid, so try it now while it is $0.

Other notable updates include the following:

• Ventoy has been updated to version 1.1.03 with fixes for Windows 11 installation issues.
• Copilot in Excel can now reference files from Word, PowerPoint, and other applications.
• Microsoft announced the Majorana 1 chip, a breakthrough development in quantum computing.

And here are the newest drivers and firmware released this week:

• Nvidia 572.47 WHQL with RTX 5070 Ti support and DLSS 4 for Marvel Rivals.
• The Surface Laptop 5, Surface Go 3, and Surface Book 3 received new firmware with security patches.

Reviews are in

Here is hardware and software we reviewed this week

This week, Robbie Khan dropped a review of the Ducky One X, the world's first inductive switch keyboard tailored for typing and gaming. There is a lot to love about this keyboard, but it also has some cons you should consider before spending $179.

On the gaming side

Learn about upcoming game releases, Xbox rumors, new hardware, software updates, freebies, deals, discounts and more.

Microsoft unveiled a new pack of games for Xbox and PC Game Pass subscribers. In the second half of February 2025, Avowed, EA Sports F1 24, Warhammer 40,000: Rogue Trader, and Watch Dogs: Legion will be available. Some games are also leaving the service, so check out the full list here.

Rare unveiled Season 15 for Sea of Thieves. The latest content drop includes some seriously cool stuff, like ancient megalodons with superpowers, new wildlife, and a lot more. The update arrived on February 20 and is available on PC, Xbox consoles, and PlayStation 5.

Indiana Jones and the Great Circle received a big update with a few important changes. The recently launched title now supports DLSS 4, AMD's frame-generation tech, and an absolutely massive list of fixes for gameplay, story, and more.

Rockstar is finally ready to deliver the long-promised GTA V console upgrade to PC. On March 4, the 12-year-old game will receive various changes, such as new vehicles and animals, improved career tracking, graphical updates, FSR 3 and DLSS 3 support, faster loading with DirectStorage, and more. Note that the update has its own hardware requirement, which you can check out here.

Xbox consoles received a new software update, which adds a rather unconventional feature. Now, Xbox users can connect absolutely massive hard drives to their systems. How massive? Think of larger than 16TB. Previously, 16TB was the maximum, but now, you can connect drives like the discounted WD Red Pro 24TB NAS HDD.

The GeForce NOW streaming service received a new load of games that you can play in the cloud. They include Awoved, Lost Records, Warhammer: Rogue Trader, HUMANITY, and more. Overall, this week, Nvidia added six new titles to its cloud streaming service.

Deals and Freebies

This week's Weekend PC Game Deals features several giveaways from the Epic Games Store on PC and mobile, plus a big drop of various discounts across stores like Steam, Humble Bundle, and more.

Other gaming news include the following:

• Kingdom Come: Deliverance 2 sold two million copies in just two weeks.
• Bethesda announced a charity auction, and the winner will become an NPC in The Elder Scrolls VI.
• Roller Coaster Tycoon 3 is finally coming to Xbox and PlayStation.

• Black Ops 6 gains a TMNT crossover, new maps, weapons, and more soon.

• Microsoft unveils Muse AI model for game prototyping and preservation.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

As reported by Mark Gurman in Bloomberg's Power On newsletter, Apple is gearing up to unveil the 13-inch and 15-inch MacBook Air with M4 chips as soon as March. The company is reportedly preparing its marketing, sales, and retail teams for the product release.

Additionally, the inventory of older MacBook Air models is running low in the Apple Store, a subtle sign of Apple preparing to release a new product model.

"I'm told the company is now gearing up for a release in March and has started preparing its marketing, sales and retail teams for the debut. The current models are also seeing inventory being drawn down at stores — another sign the launch is nearing."

The Apple M4 chip powers some of the company's best-selling products, including the new Mac mini, iPad Pro, iMac, and MacBook Pro. As for the MacBook Air with M4, the device has already appeared in Geekbench with 24GB of unified RAM and a base frequency of 4.41 GHz.

Geekbench Metal, which evaluates GPU performance, has awarded a score of 54,806 to the M4-powered MacBook Air. For comparison, the MacBook Air with M3 chips could score 41,045 in the same test.

But here comes the surprising part: the M4 MacBook Air's performance in Geekbench Metal is a mere 5 percent lower than the MacBook Pro M4, which boasts a Metal score of 57,603. Considering the price disparity between the two models, the M4-powered MacBook Air has done an impressive job on the GPU side.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

However, there is a method to keep using PhysX on the newest hardware and not lose performance: pairing a previous generation Nvidia GPU to just run the PhysX elements. That's exactly what Reddit user "jerubedo" has opted for, letting an RTX 3050 run the proprietary physics effects while a 5090 runs the game.

The user posted on the Nvidia subreddit that they purchased the RTX 3050 for this use because they still play some classic games from time to time and clearly wanted the extra effects without the loss of performance.

Some useful benchmarks were provided by "jerubedo" from games running on this RTX 5090 and RTX 3050 machine too, featuring some massive jumps in performance on Mafia II classic, Batman Arkham Asylum, and others:

Mafia II Classic results:

 

• Benchmark run without the 3050 and max settings: 28.8 FPS
• Benchmark run with the 3050 and max settings: 157.1 FPS

 

Batman Arkham Asylum results:

 

• Benchmark run without the 3050 and max settings: 61 FPS (but with MANY of the scenes in the low 30s and 40s)
• Benchmark run with the 3050 and max settings: 390 FPS

 

Borderlands 2 results:

 

• 1 minute gameplay run in area with heavy PhysX without the 3050 and max settings: Could not enable PhysX at ALL. I tried everything including different legacy versions of PhysX and editing .ini files, all to no avail.
• 1 minute gameplay run in area with heavy PhysX with the 3050 and max settings: 122 FPS

 

Assassin's Creed IV: Black Flag results:

 

• Playthrough of intro without 3050 at max settings: 62 FPS (engine locked).
• Playthrough of intro with the 3050 at max settings: also 62 FPS (engine locked).
• It seemed PhysX wasn't dragging this title down when using the CPU for PhysX. I saw the effects working as pieces of the ship were splintering off into the air as it was being hit by cannon balls.

Jerubedo added later that the RTX 3050 usage stays around 40% to 50% while it's doing PhysX work, peaking at 60%, and expressed doubts about older cards being able to get the same level of performance.

For those who want to take the same route, the Nvidia Control Panel currently offers the option to set a card as the main PhysX processor. This setting can be found in the 3D Settings section of the panel, inside "Configure Surround, PhysX" category. Here, the PhysX settings box can be set to a specific GPU before ticking the "Dedicate to PhysX" box to complete the process.

While PhysX is still supported on Nvidia cards prior to the RTX 50 series, the RTX 3050 had been a good option for this use case because of its tiny form factor, depending on the variant. Some versions can run off the PCIe slot's power without external power cables as well, making it a low-profile option for those with extra PCIe slots.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

Paul's Hardware (1.51M subscribers)

February 23, 2025

Video length: 13m 31s

0:00 Welcome to Paul’s Tech News - Feb 23, 2025

1:54 Some RTX 5090 GPUs Are Missing ROPs, Which Drops Performance

4:04 Anti-melting 12VHPWR adapter has active cooling and temp monitoring

5:13 NVIDIA now investigating black screen issues across RTX Series GPUs

6:17 Radeon RX 9070 and RX 9070 XT Launch Approaches, No Ref Design, More Leaks

8:43 5070 Ti launch: Almost OK, if it was actually $750

TECH BRIEFS

10:29 NVIDIA launches Verified Priority Access for GeForce RTX 5090 and 5080

11:26 PC Hardware Prices Continue to Rise Due to Tariffs; 25% Chips Tariff Threatened

12:17 Nvidia's Singapore GPU sales are 28% of its revenue, but only 1% go to the country

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

In a statement to The Verge, Nvidia GeForce global PR director Ben Berraondo said that this is a "rare issue affecting less than 0.5%" of its manufactured GeForce RTX 5090/5090D and 5070 Ti cards and urged affected owners to contact the manufacturer for a replacement.

Here's the full statement:

We have identified a rare issue affecting less than 0.5% (half a percent) of GeForce RTX 5090 / 5090D and 5070 Ti GPUs which have one fewer ROP than specified. The average graphical performance impact is 4%, with no impact on AI and Compute workloads. Affected consumers can contact the board manufacturer for a replacement. The production anomaly has been corrected.

First reported by TechPowerUp, the publication quickly found reports of the anomaly across Zotac, MSI, Manli, Gigabyte, Palit, and Inno3D. While the original report only found the issue on RTX 5090 and RTX 5090D cards, but Nvidia says that even the 5070 Ti is affected too. The performance hit from the missing eight ROPs can be varied, but on average it seems to make the cards perform around 4.5% slower.

Unfortunately, as this seems to be a manufacturing issue, those who are affected won't be able to fix it with a driver or BIOS update as previously hoped. Those who have RTX 5090s can check their cards for this discrepancy by booting up GPU-Z and looking at the available ROPs section, which should read 176, as listed in the specifications by Nvidia. On the RTX 5070 Ti, it should read 96 ROPs, while defective units will read 88 ROPs.

The latest debacle arrives as reports of melting power connectors continue to arrive from those who've been lucky enough to grab an RTX 5090. This is in addition to issues like black screens and crashes being investigated by Nvidia. Earlier today, a user on Reddit (Impossible-Weight485) posted that his new Astral RTX 5090 "caught on fire," supposedly due to blown capacitors while simply browsing the web.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

Researchers at Microsoft have announced the creation of the first “topological qubits” in a device that stores information in an exotic state of matter, in what may be a significant breakthrough for quantum computing.

At the same time, the researchers also published a paper in Nature and a “road map” for further work. The design of the Majorana 1 processor is supposed to fit up to a million qubits, which may be enough to realize many significant goals of quantum computing—such as cracking cryptographic codes and designing new drugs and materials faster.

If Microsoft’s claims pan out, the company may have leapfrogged competitors such as IBM and Google, who currently appear to be leading the race to build a quantum computer.

However, the peer-reviewed Nature paper only shows part of what the researchers have claimed, and the road map still includes many hurdles to be overcome. While the Microsoft press release shows off something that is supposed to be quantum computing hardware, we don’t have any independent confirmation of what it can do. Nevertheless, the news from Microsoft is very promising.

By now you probably have some questions. What’s a topological qubit? What’s a qubit at all, for that matter? And why do people want quantum computers in the first place?

Quantum Bits Are Hard to Build

Quantum computers were first dreamed up in the 1980s. Where an ordinary computer stores information in bits, a quantum computer stores information in quantum bits—or qubits.

An ordinary bit can have a value of 0 or 1, but a quantum bit (thanks to the laws of quantum mechanics, which govern very small particles) can have a combination of both. If you imagine an ordinary bit as an arrow that can point either up or down, a qubit is an arrow that can point in any direction (or what is called a “superposition” of up and down).

This means a quantum computer would be much faster than an ordinary computer for certain kinds of calculations—particularly some to do with unpicking codes and simulating natural systems.

So far, so good. But it turns out that building real qubits and getting information in and out of them is extremely difficult, because interactions with the outside world can destroy the delicate quantum states inside.

Researchers have tried a lot of different technologies to make qubits, using things like atoms trapped in electric fields or eddies of current swirling in superconductors.

Tiny Wires and Exotic Particles

Microsoft has taken a very different approach to build its “topological qubits.” They have used what are called Majorana particles, first theorized in 1937 by Italian physicist Ettore Majorana.

Majoranas are not naturally occurring particles like electrons or protons. Instead, they only exist inside a rare kind of material called a topological superconductor (which requires advanced material design and must be cooled down to extremely low temperatures).

Indeed, Majorana particles are so exotic they are usually only studied in universities—not used in practical applications.

The Microsoft team say they have used a pair of tiny wires, each with a Majorana particle trapped at either end, to act as a qubit. They measure the value of the qubit—expressed by means of whether an electron is in one wire or the other—using microwaves.

Braided Bits

Why has Microsoft put in all this effort? Because by swapping the positions of Majorana particles (or measuring them in a certain way), they can be “braided” so they can be measured without error and are resistant to outside interference. (This is the “topological” part of “topological qubits.”)

In theory, a quantum computer made using Majorana particles can be completely free of the qubit errors that plague other designs.

This is why Microsoft has chosen such a seemingly laborious approach. Other technologies are more prone to errors, and hundreds of physical qubits may need to be combined together to produce a single reliable “logical qubit.”

Microsoft has instead put its time and resources into developing Majorana-based qubits. While the company is late to the quantum party, it hopes to catch up quickly.

There’s Always a Catch

As always, if something sounds too good to be true, there is a catch. Even for a Majorana-based quantum computer, such as the one announced by Microsoft, one operation—known as T-gate—won’t be achievable without errors.

So the Majorana-based quantum chip is only “almost error-free.” However, correcting for T-gate errors is much simpler than the general error correction of other quantum platforms.

What now? Microsoft will try to move ahead with its road map, steadily building larger and larger collections of qubits.

The scientific community will closely watch how Microsoft’s quantum computing processors operate, and how they perform in comparison to the other already established quantum computing processors.

At the same time, research into the exotic and obscure behavior of Majorana particles will continue at universities around the globe.

Source

Hope you enjoyed this news post.

Thank you for appreciating my time and effort posting news every day for many years.

News posts... 2023: 5,800+ | 2024: 5,700+ | 2025 (till end of January): 487

RIP Matrix | Farewell my friend  

When you cover tech news for over a decade, you notice news often comes in bunches. This past week, the collective tech world must have been feeling a bit nostalgic, because Windows Phone and Windows 95 were both in the headlines. We also saw the Halo TV series potentially get a second life through Netflix. And while not necessarily good news, our Senior Editor Zac Bowden broke down how Microsoft's Surface lineup has changed over the years.

Surface is dead... kind of