Latest articles, news and analysis in the GPU category.
Apple is exploring a strategic shift in its chip production, potentially using Intel’s foundry for next-generation silicon, which could reshape supply chains and transistor density competition while maintaining TSMC as a key partner.
Two new ASUS ProArt OLED monitors deliver 10-bit color depth, 99% DCI-P3 coverage, and up to 1000 nits peak brightness—positioning them as premium options for video editing and design work.
A recent AMD driver update introduces compatibility issues with Blender's Cycles path-tracing engine. Users experience crashes due to a ROCm runtime mismatch, prompting a recommendation to revert to an older driver version until full support is restored.
A new compression layer from AMD promises to shrink game geometry files by up to 22%, but it won’t speed up in-memory processing—and that could matter more for performance.
Gigabyte introduces a limited-time bundle pairing select OLED monitors with three months of Xbox Game Pass Ultimate, emphasizing high refresh rates and immersive visuals for diverse gaming experiences.
NVIDIA's upcoming GPU architecture, codenamed Vera Rubin, is designed specifically for AI workloads and expected to arrive in July. With potential performance gains in matrix operations and a focus on large-scale neural networks, this chip could redefine the landscape of AI acceleration, though supp...
The NVIDIA V100, an architecture from 2017, is now available at a fraction of its original price—yet it remains a dominant force in AI workloads, outpacing modern consumer GPUs in benchmarks. This revival raises questions about platform compatibility and long-term supply for IT teams.
The founder of SMIC argues that the industry's fixation on ultra-fine nodes like 3nm may be misguided, advocating instead for greater emphasis on efficiency gains at larger nodes and packaging innovations.
A Gigabyte RTX 5060 Ti graphics card has surfaced with AMD branding, suggesting a potential mix-up in manufacturing or packaging. The incident highlights the complexities of GPU supply chains and the challenges for creators seeking consistent hardware experiences.
AMD's Dense Geometry Format (DGF) is poised to transform how AI processes complex 3D scenes, promising major efficiency gains that could reshape ray-traced applications and future RDNA GPUs.
The latest update to the PC specs for 007: First Light shows a significant shift in hardware priorities, with an emphasis on raw performance over advanced graphical features like path tracing. This change raises questions about pricing and availability, particularly for data-intensive workloads.
The RTX 50 series delivers unprecedented performance, but its steep pricing and unproven efficiency could leave PC builders on the fence. Is this a must-have upgrade or a high-stakes gamble?
AMD's latest RDNA 4 GPUs, including the RX 9070 XT and RX 9060 XT, are now selling below MSRP in China, marking a rare pricing correction. This shift could signal broader supply adjustments for gamers worldwide.
AMD's Radeon RX 9070 XT and RX 7650 GRE graphics cards have seen a sharp price drop in China, reflecting broader market dynamics. The move highlights the impact of inflated pricing on consumer demand and the challenges of maintaining competitive positioning.
The new Gigabyte RTX 5090 Infinity redefines cooling efficiency with a triple-fan setup disguised as dual, while delivering a factory overclock of 323 MHz. This article explores the practical implications for enterprise buyers and the tradeoffs in performance and heat management.
The AORUS GeForce RTX 5090 INFINITY 32G introduces a new paradigm in high-end GPU design, blending award-winning aesthetics with adaptive cooling technology that prioritizes efficiency without sacrificing performance or thermal management.
Gamers are split on NVIDIA's latest AI upscaling technology, with a majority preferring original visuals over photorealistic enhancements. Real-world results will shape opinions before the fall launch.
The Black Magic Design Puck, with its 16GB GDDR5 RAM and 8GB HBM2 storage, is poised to change how developers approach platform lock-in. Its unique architecture demands careful consideration for upgrade paths and buying timing.
