João Silva

3 days ago | kitguru.net | João Silva

Nvidia scaled back its Computex presence last year but the company is looking to ramp things back up for Computex 2025. Nvidia has now announced its keynote, taking place on the 18th of May. Nvidia's Computex keynote is scheduled for May 18th at 8 p.m. PDT/May 19th at 4 a.m. BST. While TAITRA, the organisers of Computex, haven't explicitly detailed the keynote's content, the official description strongly suggests that Artificial Intelligence will be the central theme of the presentation.

3 days ago | kitguru.net | João Silva

Avalanche Software has released a new patch for Hogwarts Legacy, introducing official support for AMD FSR 3 and Intel XeSS 2.0. This update enables both AMD and Intel GPU users to use frame generation, potentially leading to significant performance gains. As part of the integration of Intel XeSS 2.0, the April 17th update brings integration for Intel's Low Latency (XeLL) and Frame Generation (XeFG) technologies.

3 days ago | kitguru.net | João Silva

Researchers at Fudan University have developed a new non-volatile flash memory technology that could revolutionise storage. Named “PoX”, this flash memory could potentially bridge the performance gap between volatile and non-volatile memory, paving the way for advancements in data-intensive AI systems. Traditional volatile memory solutions, such as SRAM and DRAM, offer fast write speeds (1-10 nanoseconds) but lose data when power is lost.

5 days ago | kitguru.net | João Silva

A recent leak suggests that AMD might be using an older architecture for one of the Ryzen Z2 series APUs. The leak claims that the Ryzen Z2 A might be using the Van Gogh architecture, the same architecture found in the Steam Deck APU. The Ryzen Z2 series already features various architectures, including Rembrandt, Hawk Point, and Strix Point. AMD is reportedly planning to add two more: the Ryzen Z2 A and the Ryzen AI Z2 Extreme.

5 days ago | kitguru.net | João Silva

Intel was among the first GPU vendors to introduce dedicated generative AI software. Their AI Playground software, optimised for Intel Arc GPUs and integrated graphics solutions is described as an “AI HUB,” as it can run chat-based Large Language Models (LLMs) in addition to generating images. While widely praised for its user-friendliness, AI Playground's limitation to Intel hardware has been a drawback for some. However, the accessibility of AI Playground might be expanding soon.