|
|
|
www.design-reuse-embedded.com |
|
Nvidia reportedly using TSMC N5 node for GeForce RTX 4000 Series
pcinvasion.com, Nov. 30, 2021 –
Nvidia is reportedly planning to introduce two new GPU architectures to its lineup of RTX 4000 Series cards. Making use of TSMC's N5 process technology, these new architectures will net significant performance boosts in comparison to the current node used for the RTX 3000 Series lineup. This report comes from DigiTimes.
Nvidia uses Samsung's 8LPP fabrication process to make its RTX 3000 Series GPUs. The Ampere architecture is also already making use of TSMC's capabilities with the N7 node to produce the high-performance GA100 GPU for professional applications. But now, Nvidia has reportedly decided to use TSMC's N5 process technology for its lineup of RTX 4000 Series cards, cutting out Samsung's role in the manufacturing process.
In terms of logistics and partner agreements, it makes sense for Nvidia to streamline the production of its next GPU lineup somewhat. The dual foundry approach to the gaming and professional lineups makes sense for the company in theory, but there have been concerns in regards to Samsung's ability to provide adequate yields in sufficient volume on its own advanced nodes for future generations of GPUs. With this in mind and the proven track record of TSMC's N5 node, this report sounds more than believable. Nvidia has also worked with TSMC extensively in the past to produce previous generations of GPUs.
As for the new architectures, Lovelace and Hopper will be designed to appeal to all corners of the market. Lovelace will be used for GPUs geared towards gaming, while Hopper will be the architecture used for data centers and other tasks besides gaming.
Back
Contact Us
Partnership Offers