Gaming PC

AMD RDNA 3 GPU Unveil “together we advance_gaming” Live Blog (1pm PT/20:00 UTC)

Following AMD’s major CPU launches this year (Zen 4 architecture and Ryzen 7000 family), today AMD is showing off their GPU architectures in the sun at their gaming-focused ‘together we Advance_gaming’ event. Today’s event is all about AMD’s next-generation Radeon GPU architecture, RDNA 3, which promises, among other things, a 50% performance-per-watt improvement over his RDNA 2 (Radeon RX 6000 series) parts of the previous generation. It has been.

Notably, unlike AMD’s CPU event in August, AMD kept their (video) cards close to their chests for this event. We know it’s all about the RDNA 3 architecture, but AMD is silent on product information and details. For example, the “Radeon RX 7000 Series” branding has been completely removed from all official AMD communiqués. So while the Radeon RX 7000 series is still the branding we’d expect, AMD leaves plenty of room to throw in a few curveballs here.

In any case, AMD has provided some high-level details about the RDNA 3 architecture over the course of the year. The biggest items revealed so far are that AMD is targeting another 50% performance-per-watt boost and that these new GPUs (Navi 3x) will be on the 5nm process (definitely TSMC). to be manufactured. Beyond that, AMD hasn’t given any guidance on what to expect in terms of performance.

One interesting thing, however, is that AMD has confirmed that they will be adopting chiplets in this generation of products. To what extent, and if it is in all parts or just some, remains to be seen. It is in some ways the holy grail of building GPUs, as it provides options for scaling up your GPU. That said, the sheer amount of data (on the order of terabytes per second) that must be passed between different parts of the GPU is very difficult to handle, and very difficult when multiple operations are required. This is also the holy grail because it is necessary. Allows the chip GPU to appear as a single device.

It also looks like AMD is planning an even more significant upgrade to its overall GPU architecture. It’s still unclear what exactly “redesigned compute unit” and “optimized graphics pipeline” mean.

The answers to all of these, and more, will appear here in a few hours. Join us on November 3rd.rd, at 1:00 PM PT (20:00 PT) for full live blog coverage of AMD’s latest GPU announcements. I can’t wait!

Related Articles

Back to top button