-
Robot
Most Wanted
- Rep Power
- 82
AMD Previews 432 GB HBM4 Instinct MI400 GPUs and Helios Rack‑Scale AI
At its "Advancing AI 2025" event, AMD rolled out its new Instinct MI350 lineup on the CDNA 4 architecture and teased the upcoming UDNA-based AI accelerator. True to its roughly one‑year refresh rhythm, the company confirmed that the Instinct MI400 series will land in early 2026, showcasing a huge leap in memory, interconnect bandwidth, and raw compute power. Each MI400 card features twelve HBM4 stacks, providing a whopping 432 GB of on-package memory and pushing nearly 19.6 TB/s of memory bandwidth. Those early HBM4 modules deliver approximately 1.6 TB/s each, just shy of the 2 TB/s mark. On the compute front, AMD pegs the MI400 at 20 PetaFLOPS of FP8 throughput and 40 PetaFLOPS of FP4, doubling the sparse-matrix performance of today's MI355X cards. But the real game‑changer is how AMD is scaling those GPUs. Until now, you could connect up to eight cards via Infinity Fabric, and anything beyond that had to go over Ethernet.
The MI400's upgraded fabric link now offers 300 GB/s, nearly twice the speed of the MI350 series, allowing you to build full-rack clusters without relying on slower networks. That upgrade paves the way for "Helios," AMD's fully integrated AI rack solution. It combines upcoming EPYC "Venice" CPUs with MI400 GPUs and trim-to-fit networking gear, offering a turnkey setup for data center operators. AMD didn't shy away from comparisons, either. A Helios rack with 72 MI400 cards delivers approximately 3.1 ExaFLOPS of tensor performance and 31 TB of HBM4 memory. NVIDIA's Vera Rubin system, slated to feature 72 GPUs and 288 GB of memory each, is expected to achieve around 3.6 ExaFLOPS, with AMD's capabilities surpassing it in both bandwidth and capacity. And if that's not enough, whispers of a beefed‑up MI450X IF128 system are already swirling. Due in late 2026, it would directly link 128 GPUs with Infinity Fabric at 1.8 TB/s bidirectional per device, unlocking truly massive rack-scale AI clusters.

More...
-
Robot
Most Wanted
- Rep Power
- 82
-
Robot
Most Wanted
- Rep Power
- 82
NVIDIA Prepares Cut-Down GeForce RTX "5090DD" for China
When we heard the news that NVIDIA will be halting shipments to Chinese customers for its GeForce RTX 5090D GPU in Q2, there were some speculations that the company is preparing to make a comeback. Leaving out the Chinese market is a tough decision from a business standpoint, but NVIDIA complies with export regulations imposed by the US administration under Donald Trump. However, the company now plans to keep its presence in China with the release of the export-abiding GeForce "RTX 5090DD" GPU. Carrying a GB202-240 die, it will be a cut-down version with reportedly only 14,080 CUDA cores present, paired with 24 GB of GDDR7 memory.
Chinese gamers will notice a significant difference, but so will anyone trying to acquire these GPUs for non-gaming purposes. As the US administration has banned shipments of high-end chips to China to stop them from training military-grade AI, gaming GPUs have been a target of AI firms. Now, a modified GeForce RTX 5090 DD, which not only cuts with raw compute but also firmware, should be sufficient for gaming. Carrying a PG145 SKU 40 board, the new RTX 5090DD GPU will also feature a modified 384-bit bus for its 24 GB of memory, down from the original RTX 5090D's 512-bit bus.
Update 09:45 UTC: According to one of the most accurate NVIDIA leakers, kopite7kimi, the core count for this GPU is supposed to be 21,760 CUDA cores, with 575 W TDP. The leaker did note that there is some surprise, so NVIDIA has definitely changed something big under the hood to comply with export regulations.

More...
-
Robot
Most Wanted
- Rep Power
- 82
(PR) AMD Spartan UltraScale+ FPGAs Begin Production Shipments
AMD is pleased to announce that the initial devices in the Spartan UltraScale+ cost-optimized family are now in volume production! The three smallest devices, the SU10P, SU25P, and SU35P, are currently available for order with production device support in AMD Vivado Design Suite 2025.1. Built for cost-sensitive edge applications requiring high I/O, low power, and state-of-the-art security features, this new offering in the AMD Cost-Optimized Portfolio brings modernized connectivity, post-quantum cryptography, and more to the proven UltraScale+ portfolio of FPGAs and adaptive SoCs. First production shipments of the three lowest-density devices mark a significant milestone in the availability of proven, small FPGA solutions for the low and mid-range markets.
A Proven Foundation for Fast Time to Market
Engineers looking for cost-optimized, compact FPGAs also want simplicity and a fast time to market. They need programmable logic, high I/O, robust security, and world-class reliability. They also need tools that deliver push-button timing closure and rapid debug capabilities. Above all, they need a low-risk path to market. Built by the industry leader using proven UltraScale+ technology, Spartan UltraScale+ FPGAs deliver exactly that.
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
ASUS RTX 5090 Dhahab Edition Raises $24,200 in Charity Auction
ASUS has recently held a charity auction with special edition ROG Astral GeForce RTX 5090 Dhahab Edition graphics card signed by NVIDIA's CEO, Jensen Huang. ASUS managed to raise $24,200, which will go to the TZU CHI USA Foundation, a volunteer-based non-governmental organization providing charitable assistance, medical aid, disaster relief, and educational services, and environmental protection programs to those in need.
The auction period lasted from June 4 to June 13, and showing how tight the race was, the three highest bidders were just $10 apart. As the ROG Astral GeForce RTX 5090 Dhahab OC Edition is quite a rare graphics card and sold exclusively in the Middle East, the one signed by Jensen Huang is quite a catch. Hopefully, it won't end up on eBay.
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
Microsoft Breaks Windows Hello: Darkness Is Now Its Enemy
Microsoft has updated Windows Hello facial recognition to require a compatible color camera, in addition to infrared sensors, for user sign-in. Although this change was announced in April's Patch Tuesday release (KB5055523), many users only noticed the impact when their devices stopped authenticating in dark environments. Previously, Hello relied solely on IR sensors to capture a three-dimensional facial scan, enabling secure logins even in minimal lighting conditions. Microsoft says the new requirement addresses a spoofing vulnerability discovered earlier this year by ensuring that a visible image from the color camera accompanies the IR scan. Early reports have been mixed: some outlets confirm Hello no longer works in darkness, while others find that the glow of a laptop screen can suffice to illuminate the user's face.
A practical workaround involves disabling the color webcam in Device Manager, which forces Hello to revert to IR‑only authentication and restores dark‑room functionality. Disabling the webcam for video calls is not practical for everyone, particularly for those who need to attend mandatory daily meetings, especially in corporate settings. Fortunately, Windows Hello provides alternative sign-in options. Users can choose to sign in with a PIN, password, or fingerprint, all of which allow for quick and secure access without depending on facial recognition in low-light conditions.
More...
-
Robot
Most Wanted
- Rep Power
- 82
(PR) QNAP Unveils QSW 3000 Series 8-port Full 10GbE Lite-Managed Switc
QNAP Systems, Inc., a leading computing, networking, and storage solutions innovator, today launched its first 8-port full 10GbE Multi-Gig/Fiber Lite-managed switch, the QSW-L3208-2C6T. Featuring six 10GbE Multi-Gig RJ45 ports and two 10GbE SFP+/RJ45 combo ports, the switch also includes useful features such as LACP, QoS, and VLAN. The QSW-L3208-2C6T enables creators, small offices, and SMBs to build high-performance and reliable networks with flexible management on a budget.
"We've noticed that many creators and SMBs demand stable, high-speed networks without the complexity of full Layer 2 management," said Ronald Hsu, Product Manager of QNAP, adding "the QSW-L3208-2C6T is our first Lite-managed switch, offering just the right amount of management control for users to improve efficiency and troubleshoot quickly with minimal setup."
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
-
Robot
Most Wanted
- Rep Power
- 82
(PR) RuggON Unveils 12-inch SOL 7: The World's First Rugged Tablet Pow
RuggON, a global provider of rugged computing solutions, announces the launch of the SOL 7, a groundbreaking 12-inch fully rugged tablet and the first powered by Intel Arrow Lake processors. Designed for high-performance computing in demanding environments, the SOL 7 delivers next-generation AI capabilities, robust durability and seamless connectivity for critical sectors including public safety, automotive applications, warehousing, logistics, and agricultural.
Unmatched AI Performance with Intel Arrow Lake
Powered by the latest the latest Intel Core Ultra 5/7 processor with integrated Intel AI Boost, the SOL 7 enables powerful on-device AI for real-time analytics, image recognition, and rapid decision-making. Equipped with a suite of data capture tools including optional 2D barcode scanner with OCR, NFC reader with FIDO2 security, smart card reader, fingerprint reader and UHF RFID reader, the SOL 7 empowers field professionals to operate smarter, faster, and more securely.
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
-
Robot
Most Wanted
- Rep Power
- 82
(PR) Arteris Accelerates AI-Driven Silicon Innovation with Expanded Mu
In a market reshaped by the compute demands of AI, Arteris, Inc. (Nasdaq: AIP), a leading provider of system IP for accelerating semiconductor creation, today announced an expansion of its multi-die solution, delivering a foundational technology for rapid chiplet-based innovation. "In the chiplet era, the need for computational power increasingly exceeds what is available by traditional monolithic die designs," said K. Charles Janac, president and CEO of Arteris. "Arteris is leading the transition into the chiplet era with standards-based, automated and silicon-proven solutions that enable seamless integration across IP cores, chiplets, and SoCs."
Moore's Law, predicting the doubling of transistor count on a chip every two years, is slowing down. As the semiconductor industry accelerates efforts to increase performance and efficiency, especially driven by AI workloads, architectural innovation through multi-die systems has become critical. Arteris' expanded multi-die solution addresses this shift with a suite of enhanced technologies that are purpose-built for scalable and faster time-to-silicon, high-performance computing, and automotive-grade mission-critical designs.
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
NVIDIA GeForce 576.80 Game Ready WHQL Drivers Released
NVIDIA has released the latest version of its GeForce Game Ready drivers, version 576.80 WHQL. The new drivers fix a couple of gaming and general bugs, and are Game Ready drivers which means these bring day-0 support for Remedy Entertainment's co-op multiplayer game, FBC: Firebreak, which includes full ray tracing, DLSS 4 with Multi Frame Generation, DLSS Frame Generation, DLSS Super Resolution, DLSS Ray Reconstruction, and NVIDIA Reflex. It also adds support for REMATCH, coming with DLSS 4 With Multi Frame Generation support, and support for DOOM: The Dark Ages' new update, which adds path tracing and DLSS Ray Reconstruction. The new drivers also add support for 46 new and updated NVIDIA app DLSS overrides, and 4 new Optimal Playable Settings profiles.
NVIDIA also fixed several gaming and general bugs, including stability issues in games like Dune: Awakening, EA Sports FC 25, Clair Obscur: Expedition 33, Marvel Rivals, Gray Zone Warfare, Forza Horizon 5, and others, a well as some general bugs including GeForce RTX 5090 FE acoustic improvements for idle and low loads, and fix for an issue where certain DisplayPort 1.4 monitors may display random black screen flicker when connected to DisplayPort 2.1 graphics card.
DOWNLOAD: NVIDIA GeForce 576.80 Game Ready WHQL
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
(PR) Gigabyte Announces the Availability of MO27Q2A
GIGABYTE, the world's leading computer brand, announces the availability of the MO27Q2A QD-OLED gaming monitor, the brand's first QHD 280 Hz OLED display. Featuring a blazing-fast 280 Hz refresh rate, 0.03 ms GtG response time, and VESA ClearMR 15000 certification, the MO27Q2A delivers ultra-smooth motion and competitive speed for next-gen gaming.
Built with a 27" 2560×1440 OLED panel, the MO27Q2A supports DisplayHDR True Black 400, offers 99% DCI-P3 color gamut, Delta E
-
Robot
Most Wanted
- Rep Power
- 82
(PR) TerraMaster D4 SSD Unveiled: 40 Gbps All-Flash 4-Bay Storage Solu
TerraMaster, a leading innovator in storage solutions, proudly announces the launch of D4 SSD, a 4-bay all-flash direct-attached storage (DAS) enclosure designed for media creators, filmmakers, and professionals seeking unparalleled speed and flexibility. With USB4 support delivering up to 40 Gbps bandwidth, the D4 SSD offers exceptional performance and seamless integration with NAS systems, including the forthcoming TerraMaster F4 SSD All-Flash NAS, easily expand the storage space of NAS to meet users' diverse storage needs.
The D4 SSD utilize the USB4 protocol with support for up to 40 Gbps data transfer bandwidth. When equipped with four M.2 SSDs (such as Samsung 990PRO 4 TB) and configured as RAID 0 via macOS Disk Utility, it achieves read/write speeds of up to 3224 MB/s. In single-disk operation, read speeds can reach 1608 MB/s. Fully compatible with Thunderbolt 3/4/5 and USB4/3.2/3.1/3.0 protocols, the D4 SSD offers exceptional cross-platform compatibility.
Read full story
More...
-
Robot
Most Wanted
- Rep Power
- 82
(PR) Marvell Develops Industry's First 2 nm Custom SRAM
Marvell Technology, Inc., a leader in data infrastructure semiconductor solutions expanded its custom technology platform with the launch of the industry's first 2 nm custom Static Random Access Memory (SRAM), designed to boost the performance of custom XPUs and devices powering cloud data centers and AI clusters. Combining advanced custom circuitry and software from Marvell with core SRAM and cutting-edge 2 nm process technology, Marvell custom SRAM delivers up to 6 gigabits of high-speed memory while significantly reducing memory power consumption and die area at comparable densities.
Custom SRAM is the latest Marvell innovation designed to enhance memory hierarchy performance within accelerated infrastructure. Marvell previously introduced its CXL technology for integration into custom silicon to add terabytes of memory and supplemental compute capacity to cloud servers and unveiled custom HBM technology that increases memory capacity by up to 33% while reducing the space and power required for dense high-bandwidth memory (HBM) stacks inside XPUs.
Read full story
More...
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules