One noteworthy aspect of the Titan 360 RX RGB cooler is its exceptionally low operating noise. In regular usage scenarios, this cooler generates lower sound levels compared to many similar products available today, enhancing its suitability for quiet computing environments. Corsair’s integration of the new RX-series cooling fans also contributes significantly to both noise reduction and thermal efficiency. Furthermore, the single-cable design simplifies installation, creating a clean aesthetic and straightforward user experience.
» Read Review

EA’s latest release, F1 25, has its download schedule and PC specs out—here’s what gurus need to know. If you’ve pre-ordered the Iconic Edition, you can start downloading on May 27, 2025, at 00:00 UTC across Steam, the EA App, PS5, and Xbox Series X|S. Standard Edition pre-load kicks off two days later, May 29, at the same midnight UTC mark. EA Play members get a five-hour sneak peek of Standard Edition on May 27, while EA Play Pro members get full Iconic Edition access right away. This staggered approach helps keep things smooth for everyone logging in on launch day. On the PC side, there are three tiers of recommended hardware so you can gauge what your rig needs. The entry-level setup calls for an Intel Core i5-8400 or AMD Ryzen 5 2600 CPU, 8 GB of RAM, and a GeForce GTX 1060 or Radeon RX 580 GPU, plus 100 GB of SSD space. If you’re aiming for higher fidelity, the mid-range build steps up to an Intel Core i7-9700K or Ryzen 7 3700X, 16 GB RAM, and an RTX 2070 Super or RX 6800. For those chasing every visual enhancement, the Ultra Max configuration demands an Intel Core i9-10900K or Ryzen 9 5900X, 32 GB RAM, and an RTX 3080 Ti or RX 7900 XT, again with 100 GB on an SSD. That top-end spec unlocks all the advanced features, including path tracing.
» Read full story

VIPkeysale brings you an exclusive deal you can’t miss! Buy Windows key for just $13 and upgrade to Microsoft Office for only $26. Take advantage of this limited-time opportunity to get genuine Microsoft software at a fraction of the price. Don’t wait — these prices won’t be around forever! All of the prices that you see below are now available to anyone who enters the discount code GURU30. Let’s dive in!

Microsoft Windows 11 Pro OEM CD-KEY GLOBAL 22.0USD--(after 30% discount GURU30)
Microsoft Windows 11 Home OEM CD-KEY GLOBAL 22.7USD--(after 30% discount GURU30)
Microsoft Windows 10 Pro OEM CD-KEY GLOBAL 15.8USD--(after 30% discount GURU30)
Office2016 Professional Plus CD Key Global 26.1USD--(after 30% discount GURU30)
Office2019 Professional Plus CD Key Global 45.3USD--(after 30% discount GURU30)

Next Level Racing’s new F-GT Pro cockpit hits all the right notes for sim racing gurus. Its 50 mm tubular steel frame and reinforced brace keep things solid, and switching from a Formula vibe to a GT setup is just a matter of moving an adjustable pedal plate—no wrenches required. The adjustable backrest slides smoothly to fit gurus from about 155 cm up to 200 cm tall, and it’s rated for payloads up to 150 kg. Despite packing in all that rigidity, the main frame only tips the scales at about 49.5 kg, and it measures 542 mm wide by 1,667 mm long by 946 mm high, so most home setups will have plenty of space. What really sets this cockpit apart is how you can expand it. Three bolt-on modules are available right away: a direct-mount arm for your display, a slide-out tray for a keyboard and mouse, and a shifter plus handbrake arm combo. Each piece locks into place with captive fasteners so you can dial in the perfect control positions. The monitor arm and keyboard tray each add around 300 mm to the cockpit’s footprint, giving you plenty of ergonomic range. And yes, it all meshes seamlessly with other Next Level Racing gear—motion platforms, secondary monitor brackets, you name it.
» Read full story

DeepCool’s CH690 DIGITAL is a mid-tower PC chassis that aims to blend component visibility with technical control. It comes in black and white, each featuring 2.5-sided tempered glass panels—front, left and part of the rear—for an unobstructed view of the hardware inside. The rear glass can be swapped out for a mesh panel and fan bracket if you need extra airflow. The black model goes for roughly USD 133, and the white one is about USD 148. Both hit the market on May 16, 2025, with a one-year warranty standard. What sets this case apart is its digital readout on the lower left chamber. Instead of opening an app, gurus can glance at real-time system data: temperatures, power draw, utilization rates and clock speeds. Flip between CPU mode to check fan RPM or PSU mode to see additional fan stats. It’s a straightforward way to monitor performance without juggling software windows.
» Read full story

You’ve probably heard about Phison’s new PS5028-E28, the next-generation PCIe 5.0 SSD controller that’s stirring up some excitement in storage circles. In a recent preview from Toms Hardware, the E28 went head-to-head with three heavy hitters: Samsung’s 9100 Pro (which uses the Presto controller), Micron’s 4600, and Acer’s Predator GM9000—both of the latter on the SMI SM2508 platform. The benchmark of choice was PCMark 10’s Full System Drive test, a suite that mimics everyday PC tasks like opening apps, firing up games, and moving files of various types. It’s a solid way to see how drives perform under normal desktop use rather than just pushing them to synthetic extremes. Here’s the gist: the Phison E28 crushed the competition in this test. It scored 7,978 points, while the 9100 Pro trailed at 6,767, the Micron 4600 at 6,748, and the Predator GM9000 at 6,040. When you look at sustained bandwidth, the E28 managed about 1.21 gigabytes per second. That’s roughly 16 percent better than the Samsung and Micron, and about 29 percent more than the Predator. Latency—which you want as low as possible—was 20 microseconds on the E28, compared to 24 microseconds on both the 9100 Pro and 4600, and 27 microseconds on the GM9000. In short, the E28 is quicker at getting data back to you, which can really help when you’re juggling multiple tasks or running high-I/O workloads.
» Read full story

First off, Amazon accidentally spilled the beans on XFX’s tweaked RX 9060 XT models, showing a boost speed cranked up to 3 320 MHz—about 90 MHz faster than AMD’s rumored base spec for both the 16 GB and 8 GB editions. Then, momomo_us dropped a more detailed leak with two clock profiles: the top-end version hits a 3 320 MHz boost with a steady 2 780 MHz “game clock,” while the mid-tier variant offers a 3 290 MHz boost alongside a 2 700 MHz sustained speed. Those numbers even top the RX 9070’s 3 100 MHz boost on its fastest official cards. The Navi 44 chip at the heart of the RX 9060 XT has been rumored to run hotter and faster than the bigger Navi 48. In simple terms, the “game clock” is the floor—the speed you get during long gaming sessions—while the “boost clock” is the ceiling for short bursts. By splitting clock targets like this, board partners can offer different “flavors” of the card—some tuned for constant performance, others pushing the absolute max for benchmarks. But the real curveball in the latest leak is the move to a full PCIe 5.0 x16 interface. Everyone expected an x8 link.
» Read full story

AI is growing, and what do we see next? Microsoft is cutting roughly 3 percent of its workforce, which equates to about 6,000 roles worldwide. The goal behind this decision is to remove overlapping management roles and simplify how teams are structured. While the majority of these cuts hit the US, teams in other regions—including those working on LinkedIn—are also part of the reduction. It’s the biggest round of layoffs Microsoft’s seen in several years. The impact reaches every division, from cloud services to software development and professional networking through LinkedIn. Microsoft hasn’t specified exactly which countries beyond the US are affected, so it’s unclear if offices in places like the Netherlands or Belgium will see any reductions. The company’s focus is on making sure resources get redirected toward high-priority projects without duplicating leadership responsibilities. In a statement, Microsoft explained that the cuts reflect “ongoing organizational changes necessary to position the company as best as possible for success in a dynamic market.” In practical terms, this means reallocating talent and budgets to areas where growth potential is highest and ensuring decision-making is more agile.
» Read full story

Microsoft pulled back the curtain on some Start menu mock-ups that didn’t make it into Windows 11’s official release. In these early sketches, designers played around with a vertical side menu, a personalized “For You” feed, and even categorizing apps by type. Imagine a left-hand panel showing your next calendar appointment and most-used files—that’s one of the prototypes they tested. Another mock-up grouped software into folders by category, but it ended up feeling too busy. Behind the scenes, Microsoft ran structured tests to see how each layout performed. They measured things like how fast participants found an app, how well they understood the menu, and overall satisfaction.
» Read full story

The ColorEdge CG3100X from EIZO is a 30.5-inch 4K display built for HDR video and color-critical work. Announced on May 13, 2025, it’s the follow-up to the CG319X that came out in 2018. What really stands out on this new model is how much brighter it goes—up to 500 cd/m² versus 350 before—and the bump in contrast to 1,800 : 1 from 1,500 : 1. You get the full DCI 4K resolution of 4,096 × 2,160, so whatever you’re grading or compositing, you’ll see it in the right dimensions. Under the hood, there’s a built-in calibration sensor that keeps your color profiles spot-on without having to attach an external meter. It handles HDR gammas for both HLG and PQ, which means you can switch between broadcast-friendly workflows and cinematic PQ grading without missing a beat. Since the sensor lives inside the monitor, you just run the calibration from the on-screen menu, and it does all the measuring and adjustments for you. On the connectivity side, it’s pretty slick. There’s DisplayPort and HDMI, but the real game-changer is the USB-C port. It pushes out up to 94 W, so you can power-charge your laptop while feeding it video and data all through one cable. There’s also a wired LAN jack for network access and a headphone jack for direct audio monitoring. It’s like having a mini docking station built right into your monitor.
» Read full story

AMD just rolled out its EPYC 4005 line of server chips, and they’re all about giving the same AM5 socket you’d use for Ryzen 7000-series CPUs a server-grade upgrade. Gurus looking to build reliable systems can now drop these into motherboards that already support consumer Ryzen 9000 parts. The big wins here are built-in ECC memory support for catching bit flips on the fly, plus official certification for enterprise operating systems. That means you can run Red Hat Enterprise Linux, SUSE Linux Enterprise Server, Ubuntu, or Windows Server without second-guessing compatibility. Let’s break down the models. You’ve got options ranging from 6 to 16 cores. The flagship EPYC 4565P is rocking 16 cores, boost clocks up to 5.7 GHz, and a 170 W TDP—right in line with what you’d expect from a top-end Ryzen 9 9950X in the desktop world. There’s also a 3D V-Cache version called the EPYC 4585PX, which tacks on an extra 64 MB of L3 cache on one chiplet for workloads that benefit from heavy cache. Both of those peak-power chips run at 170 W, while every other part in the series sticks to a 65 W envelope to keep power draw and cooling simpler. If you don’t need the extra cores or cache, you can pick from 12-, 8-, or 6-core SKUs at lower TDPs.
» Read full story

Download HWiNFO64 - a System Information and Diagnostics Tool. HWInfo64 is an application that monitors your PC's components and outs computer information about your the operating system, storage capacity, and RAM. It also displays more technical data, such as the number of logical CPUs, memory speed, and temperatures. HWiNFO is a professional hardware information and diagnostic tools supporting latest components, industry technologies, and standards. Both tools are targeted to recognize and extract the most possible amount of information about computer's hardware which makes them suitable for users searching for driver updates, computer manufacturers, system integrators and technical experts too.
» Download

Western Digital has just put money into Cerabyte, a startup that’s blending ceramic and glass to build storage media meant to last thousands of years. Instead of the usual magnetic platters or silicon flash, Cerabyte uses a ceramic–glass combo that you can literally boil in saltwater, zap with UV light, and blast with radiation—and it still holds data. The plan is to get these next-generation drives ready for the yottabyte era (that’s a trillion terabytes) so that deep-archive data stays viable for up to 5,000 years. Cerabyte kicked off some pretty wild durability tests earlier this month. They dunked a test unit in saltwater, then baked it in an oven to simulate fire and steam damage. The device survived without losing a single bit of data. They’ve also zapped it with UV to mimic sunlight exposure and subjected it to radiation levels that would fry most electronics. Even an electromagnetic pulse attack—think hardened military gear—didn’t wipe it. All of this shows the ceramic–glass substrate is basically impervious to environmental hazards that usually degrade hard drives or flash memory. The partnership with Western Digital is all about scaling up. WD brings decades of drive-design know-how and global manufacturing scale, and Cerabyte has the materials tech. They haven’t set a release date yet; instead, they’re focusing on hitting key R&D goals. First up is integrating the ceramic medium into an active archive drive with read/write heads and controllers compatible with existing interfaces. Down the line, they want to push capacity so far that storing exabytes and beyond is practical—and every bit of that data still stays intact for 5,000 years.
» Read full story

Recent filings by the Shangke Group with the Eurasian Economic Commission show that the card will ship in several versions under the Maxus label, each tagged with 8 GB of video memory. Seeing “8GB” in the model numbers gives us pretty solid proof that Nvidia is sticking to that memory size. Digging into the hardware, it looks like the RTX 5050 is built around the GB207-300-A1 GPU, mounted on a PG152-SKU50 PCB layout. There are 2,560 CUDA cores—spread over 20 SMs—driving the rendering power. On the memory side, Nvidia’s chosen a 128-bit bus populated with GDDR6 modules. That’s a step down from the GDDR6X found in higher-tier models, but it’s understandable given the card’s budget-friendly positioning. Shangke’s documents list several design flavors in the Mingxuan iCraft AIGA series. You’ll find open-air coolers for better heat dissipation in roomy cases and OTES blower coolers for tighter builds where airflow is at a premium. They even included the blower variants for the RTX 5060 and 5060 Ti in the same filing batch. Having both cooler types means system builders can pick the card that best fits their chassis constraints without sacrificing Nvidia’s feature set. The registration entries don’t specify clock speeds or power draw, so we’re still waiting on those details. However, knowing the GPU chip and memory configuration lets us estimate performance relative to the RTX 3050 series. With 2,560 CUDA cores, the RTX 5050 likely slots above the previous generation’s lower mid-range cards, offering a noticeable uplift in shader throughput and ray-tracing capability—albeit with narrower memory bandwidth due to the 128-bit interface.
» Read full story

AMD’s upcoming Zen 6 chips are shaping up to be pretty interesting. First off, these new dies will house 12 cores each instead of the eight we’ve seen up to now. That means a dual-die Ryzen 9 could rock 24 cores and 48 threads — a neat 50% jump that should seriously help when gurus run heavy multi-threaded apps or stream and game at the same time. Cache-wise, things get even more exciting. The “Olympic Ridge” desktop parts are tipped to get 48 MB of L3 per die, compared to 32 MB right now. For server models, AMD might push that to 128 MB per die. More cache translates to snappier data access, which is great for gurus doing video editing, large-scale simulations, or running multiple virtual machines. What really caught my eye is the boost to the 3D-stacked cache. AMD’s current X3D chips stack 64 MB of extra cache on top of 32 MB base, giving you 96 MB total. Word is they’re bumping that stack to 96 MB, and when you pair it with the 48 MB base L3, you get a massive 144 MB of total cache on a single die. Paired with 12 cores, that could cut down on memory-fetch delays in cache-heavy games and applications, so gurus should see smoother frame rates and quicker load times.
» Read full story

VUSec over in the Netherlands uncovered a fresh CPU trick they’re calling “Training Solo.” If you thought Spectre Variant 2 was old news, this one brings it back to life. The gist is that attacker code can prime the indirect-branch predictor and then immediately trigger a misprediction all within the same privilege level—no jumping between user space, virtual machines, or hypervisors. That’s how it sneaks past defenses like IBPB, eIBRS, and BHI_NO, which were supposed to quarantine prediction states when switching contexts. Not every chipmaker is in the hot seat, though. AMD’s approach to branch prediction separates things in a way that this self-training attack can’t bypass. But Intel’s line-up from Cascade Lake Xeons through Cooper Lake, Whiskey Lake V, Coffee Lake Refresh, Comet Lake, Ice Lake, Tiger Lake, and Rocket Lake is squarely in the crosshairs—and that covers everything from 9th to 11th Generation Core and 2nd to 3rd Generation Xeon processors. Meanwhile, ARM’s designs share enough in common that all ARM-based cores show the same crack, so mobile and embedded systems need patches too.
» Read full story

Intel’s been busy expanding its Arc GPU lineup, and the latest filings with the Eurasian Economic Commission hint at something notable: two new Arc B580 models dubbed “iCraft 24G” and “iCraft 24G OC,” each packing a hefty 24 GB of GDDR6. That’s double the memory of the original 12 GB cards. While regulatory docs sometimes list hardware that never materializes, these entries match earlier whispers from a manufacturing partner, suggesting Intel’s seriously exploring a big-frame-buffer option for both gamers and professionals. Back when Intel launched the Arc B580 and B570, the 12 GB versions went for under USD 300—a competitive price point that challenged similar mainstream offerings. The B580 uses Intel’s BMG-G21 graphics chip, featuring up to 20 Xe cores and 2,560 shader units. That lines it up with NVIDIA’s GeForce RTX 4060, but the potential 24 GB upgrade could push it into new territory. Large memory buffers help with high-resolution textures, complex video timelines, and GPU-accelerated compute tasks. Plus, Intel’s already tipped a 24 GB Arc PRO A60 workstation card, so this might be part of a broader strategy to offer high-capacity GPUs across both consumer and pro lines. We’re just days away from Computex, where Intel typically lays out its GPU roadmap. Folks are looking to see if the 24 GB Arc B580 cards will actually launch, or if they’re still in testing. On top of that, everyone’s talking about the Arc B770, expected next quarter. That card could bring even more cores, higher clocks, and maybe another memory bump. For system integrators, boutique PC builders, and enterprise teams, these details matter: clock speeds, power consumption, memory bandwidth—all factors in gauging real-world performance.
» Read full story

The rumor mill was set in motion by a well-known forum guru named Tez2, who’s often ahead of the curve on Rockstar Games news. He claims that the team inside Rockstar has been kicking around the idea of porting GTA IV to current consoles for some time. When GTA VI got pushed back to 2026, they reportedly decided to speed up the IV remaster instead. On top of that, there’s chatter about Max Payne 3 getting a modern makeover too. The original game dropped back in 2012, and fans have been asking for a version they can load up on PlayStation 5 or Xbox Series X. From what insiders suggest, the studio might use the same base tech to bring both titles up to date. That would mean improved textures, widescreen support, and better frame rates, all running on the latest consoles and PC setups. An interesting clue came from the Liberty City Preservation Project, a fan-made mod that let you roam Liberty City inside GTA V. Rockstar and its parent company pulled that mod offline recently, which many interpret as clearing the way for an official release of GTA IV on their own terms. It makes sense: you wouldn’t want a community project stealing the spotlight just before you push your own product. So far, though, Rockstar hasn’t said a word. No press release, no teaser, no job listings mentioning “remaster work.” That could change at any moment, especially if they’re planning a surprise announcement to drum up excitement. If the remaster does drop this year, it would slot in nicely between their other big projects, giving players something to dive into while they wait for GTA VI.
» Read full story

Nvidia just bumped up the official prices on most of its graphics cards and AI accelerators, and it’s letting its partners do the same. This move came after the U.S. rolled out tighter export rules in early 2025. The key driver was shifting some of the “Blackwell” GPU production over to a TSMC plant in Arizona, which runs more expensively than the fabs back in Taiwan. Between pricier materials, logistics overhead and the higher cost of running a U.S. foundry, Nvidia felt the need to adjust its list prices so it doesn’t erode its margins. In Europe we aven’t really seen big bumps at the checkout. Distributors and retailers seem to be holding off on passing those extra costs on to buyers, at least for now. That gives gamers and workstation gurus a little breathing room before these price hikes hit the street price. Meanwhile, Nvidia is cooking up a modified version of its H100 chip, dubbed the H20, aimed at keeping its AI accelerators flowing into China under the new export limitations. The H20 follows a similar playbook to earlier H100 tweaks that skirted previous sanctions. Before all this price shuffling, Nvidia was hitting record numbers, with its data-center revenue doubling year-over-year—numbers that came in just before the most recent export clampdown. Financial analysts have been expecting Nvidia to keep its profit margins in check, even with higher production costs. The recent list-price increases look like the practical lever to pull: bump the sticker price just enough to cover the extra expenses without letting profit percentages slip.
» Read full story

The Kingston FURY Renegade G5 is a PCIe 5.0 NVMe drive built for gurus who push their machines hard. Under the hood, it uses a Silicon Motion SM2508 controller etched on a 6 nm process, plus a low-power DDR4 cache, so it stays cool and sip power even when you’re hammering through big projects or loading asset-heavy game worlds. The M.2 2280 card easily fits into modern boards and still works on older PCIe 3.0 and 4.0 slots if you’re updating an existing rig. When you peek at the numbers, the Renegade G5 clocks up to 14,800 MB/s for reads and 14,000 MB/s for writes, regardless of whether you go with the 1 TB, 2 TB, or 4 TB model. Random 4K IOPS land around 2.2 million for both reading and writing, so you won’t face storage lags when you’re editing high-res videos, running virtual machines, or tearing through massive open-world levels. On average you’ll see just 0.27 W drawing when it's idle, ramping up to around 7 W under load—but on the beefiest 4 TB it can hit 9.5 W at peak throughput. It’s not all about raw speed—Kingston engineered this drive to run reliably in all sorts of conditions. Operating temps range from 0 °C to 70 °C, and even if you ship or store the drive in extreme settings, it’s tested down to –40 °C and up to 85 °C. Vibration tolerance sits at 2.17 G when spinning and 20 G when turned off—so it handles bumps in transport—and the MTBF rating is 2 million hours. Kingston backs it up with a five-year limited warranty plus free tech support if you ever need a hand tuning your setup.
» Read full story

Download NVIDIA GeForce 576.40 WHQL Driver. In driver version 576.40 WHQL, NVIDIA has focused on correcting rendering and stability problems observed on recent hardware. Fixed bugs include random crashes in Monster Hunter Wilds, shadow flicker elimination in Dead Space Remake on RTX 50 series GPUs, and correction of color saturation at non-native resolutions. Boot failures resulting in a black display on select ASUS ROG monitors have been resolved, and minor stutter on LG televisions when using G-SYNC at lower refresh rates has been addressed. Although many issues have been fixed, some remain under review or lack conclusive root-cause confirmation. Users may still experience flickering around light sources in Ghost of Tsushima Director’s Cut, crashes in Cyberpunk 2077 when using Photo Mode with path tracing, and unexpected termination in F1 23/F1 24 at race end. EA Sports FC 25 also shows intermittent gameplay crashes, and some configurations face difficulties installing the NVIDIA Control Panel from the Microsoft Store. Notebook-specific quirks may vary by system manufacturer.
» Download

Samsung’s new Odyssey OLED G6 is a 27-inch gaming screen hitting an unheard-of 500 Hz refresh rate. It’s dropped first in Singapore, Thailand, Vietnam and Malaysia, with more regions coming online later this year. The monitor packs a QHD panel (2,560 × 1,440) built on QD-OLED tech, which means intense color reproduction and super-fast refresh for those who care about every frame. Right out of the box, the G6 sets itself apart with a 0.03 ms gray-to-gray response time. That translates to almost no ghosting in the fastest esports titles. Both NVIDIA G-SYNC compatibility and AMD FreeSync Premium Pro support are on board, so gurus with the latest graphics cards won’t run into tearing or choppy motion. On top of that, it’s got VESA DisplayHDR True Black 500 certification, delivering zoned dimming that boosts contrast—and a peak brightness level of 1,000 nits keeps highlights popping, even in bright rooms. Longevity and color precision were big focuses during development. An OLED safeguard feature cuts down on burn-in risk, which is handy for marathon gaming sessions. Pantone validation means over 2,100 colors and 110+ skin-tone shades get displayed properly, so you’re seeing content exactly as creators intended. The stand offers tilt, swivel and height tweaks to dial in perfect comfort, and a small bezel maximizes the visible area. Connectivity includes DisplayPort 1.4, HDMI 2.1 and a built-in USB hub for hooking up headsets, controllers and other gear without hunting around.
» Read full story

Here you can download the new Unified Corsair Utility Engine application (iCUE). You get to use options like configure both the lighting portions of your keyboard and mouse as well as create powerful actions such as double macros, re-mappable key assignments, timer countdowns, dpi adjustments, and more.  iCUE users can instantly synchronize RGB lighting across all their compatible Corsair products with a single click, monitor system performance from a customizable dashboard, and control hardware temperatures with in-depth fan and pump speed adjustment. At the heart of everything CORSAIR, iCUE offers intelligent control, with unlimited possibilities. With industry-leading RGB lighting control, iCUE makes it easier than ever to light up your PC. Synchronize lighting across your whole setup with just one-click, explore dozens of easy to customize lighting presets or dive into multi-layer lighting effects that mesmerize and amaze. From smooth transitions, to dynamic waves and ripples, iCUE’s lighting control goes beyond the keyboard to light up your entire setup.
» Download

The latest ROG Astral RTX 5090 Dhahab from ASUS blends high-end GPU technology with Middle Eastern design cues. Instead of sporting extra clock speed or memory, the Dhahab version is built on the exact same hardware as the regular Astral RTX 5090—same chip, same GDDR7 capacity, same thermal layout. What sets it apart is the look: a four-fan arrangement with each blade inscribed in Arabic script, an outer shell stamped with urban skyline graphics, and a layer of 6.5 grams of pure gold coating the shroud. This plating isn’t for better cooling—it’s purely decorative, aimed at builders who want a standout component in their rig. Under the hood, everything functions just like the standard Astral card. The GPU Boost speeds, power targets, and fan-curve thresholds are all identical. Gurus can tweak power-limit ceilings, push frequency offsets, and craft custom fan strategies with the included GPU Tweak III software. Voltage reads, temperature logs, and overclock profiles load in the same interface they’ve used on other ROG cards, so there’s no new learning curve.
» Read full story

AMD sixth-generation EPYC “Venice” CPUs are splitting the lineup into two main flavors: Zen 6 and Zen 6C. Think of Zen 6 as the speedster—you get higher clock speeds—and it runs on AMD’s SP7 platform. Zen 6C, on the other hand, is all about squeezing in as many cores as possible for heavy parallel jobs, running on the SP8 platform. If you’re a data-center guru looking to pick the right chip, you’ll want to match your workload needs—frequency-sensitive tasks or massive threading. Digging into the die design, each Venice chip has eight Core Complex Dies (CCDs). Picture them as clusters of mini-processors arranged around central I/O dies that handle memory and input/output traffic. For the regular Zen 6 model, each CCD packs 12 cores, giving you up to 96 cores and 192 threads in a single socket. The Zen 6C model turns up the core count to 32 cores per CCD, maxing out at 256 cores and 512 threads. Cache sizes grow in step: Zen 6 chips sport 48 MB of L3 cache per CCD (totaling 384 MB), while Zen 6C boosts that to 128 MB per CCD (1 024 MB in all). That keeps cache allocation at a steady 4 MB per core and effectively doubles what Turin could do.
» Read full story

The Amazon marketplace has spotted two new XFX Radeon RX 9060 XT graphics cards that are set to join AMD’s mid-range lineup. Both versions—the one with 8 GB of GDDR6 and the one with 16 GB—share a maximum boost speed of 3,320 MHz, but they’ve taken different approaches to keeping things cool. The 8 GB model comes with a twin-fan setup, while the 16 GB card steps it up with three fans. If you’ve seen XFX’s SWIFT look before, this should feel familiar: a dark shroud covering an aluminum fin array with vent openings at the back, though the final design might get some tweaks before launch. What really caught people’s attention are the prices showing up on Amazon. The 8 GB RX 9060 XT is tagged at USD 450, and the 16 GB version jumps to USD 520. For comparison, NVIDIA’s 16 GB RTX 5060 Ti goes for USD 429 at MSRP. Of course, Amazon often uses placeholder prices until cards actually ship, so we might see lower retail tags once AMD makes things official. But if those numbers stick, AMD will have to lean on other selling points—like higher memory bandwidth, better power draw, or extra software perks—to sway gurus who are balancing budget and performance.
» Read full story