Intel files patent for new GPU design

by

Key things

  • Intel has filed a patent for a disaggregated GPU architecture with logic chiplets for greater efficiency and modularity.
  • Potential benefits include increased computing power, scalability, and the ability to power-gate chiplets.
  • The development of this technology may lead to higher prices, but it excites tech fans and gamers.



Intel announced that it has filed a patent to develop disaggregated GPU architectures with dedicated logic chiplets, which opens up a world of possibilities by moving away from monolithic configurations and enabling independent computing processes. While Intel has provided few details about plans to implement the technology or when it will be developed, many tech fans are already considering all angles.

For several years now, Intel has been locked in fierce competition with AMD and Nvidia as it struggles to keep prices low while providing the most efficient performance possible. It hasn't been an easy road, as Intel has recently been struggling with its Raptor Lake CPUs and frustrating crashes affecting the product and its owners.


Intel Core Lunar Lake Skymont

Related

Intel Skymont details leak online

The leak reveals performance boost details and a potential launch window for Intel Lunar Lake processors that feature the new Skymont cores.

Earlier this week, Intel revealed that it has filed a patent to develop an innovative method of building disaggregated GPU architectures that use dedicated chiplets that, among other things, enable more efficient and almost completely modular GPU configurations. While modern GPUs act as a single load-sharing unit, dedicated chiplets can perform different processes independently, enabling greater computing power and scalability. But it's not all good news, as according to Wccftech, the development of such technologies often requires new and complex manufacturing processes and other technical innovations, which the company is said to be still working on. October was full of exciting news for hardware fans as Intel also unveiled its Arrow Lake processors and provided full details on the processor lineup.


Intel registers a patent for the development of discrete GPU architectures


Although Nvidia and AMD are still ahead in terms of GPU market share, many tech fans and gamers are reportedly excited about Intel's upcoming Arc Battlemage GPU architecture, which is slated for release in early 2025. A patent filing for a split GPU architecture may help turn the tide in Intel's favor and further boost that excitement, even though Battlemage GPUs will likely still be monolithic.

Intel is considered by many to be reliable when it comes to developing quality products and offering great customer support. Even so, there may be problems with the stability of the company's hardware. Recently, Intel extended the warranty on its crashing 13th and 14th generation processors in an attempt to show its commitment and willingness to fix the problem. Once the issue was resolved, Intel announced that the problem was caused by a faulty algorithm feature that caused the processors to request the wrong amount of voltage during operation.


Intel's announcement of its patent filing for a segmented GPU has many tech fans excited about the implications of this news and how this advanced technology will help revolutionize computing. Naturally, as happens with almost every innovation, there are many details that need to be fine-tuned before a discrete GPU can be manufactured, let alone sold to the public. A development effort of this magnitude usually represents a significant investment and can cause significant price increases, just as happened with the price increase between Intel's 14900KS and its 13900KS. AMD is also said to be considering the development of disaggregated GPU architectures and has filed a similar patent that could indicate the direction of the industry in the GPU market.


Leave a Comment