AMD RDNA 3 (5nm) GPUs, Radeon RX 7800 XT/7900 XT Reportedly Coming in Q2-Q3 2022

AMD’s next-gen RDNA 3 graphics cards seem to be launching sooner than expected. As per multiple sources, the higher-end Navi 33 GPU which is expected to be the flagship for the RX 7000 series will land sometime in the second or third quarter of 2022. The RDNA 3 based Radeon RX 7000 lineup should significantly improve the ray-tracing capabilities of AMD hardware, possibly with a separate pipeline for BVH and ray-tracing instructions, similar to NVIDIA’s RTX design and support for matrix instructions (Tensor core equivalent).

While AMD’s RDNA 2 RT accelerators speed up ray-box and ray-triangle intersections, NVIDIA uses the RT cores for everything from ray-box/triangle intersections to BVH traversal to ray shading. This method is much more efficient and although takes more die space, has a much lower impact on overall performance thanks to being a separate asynchronous pipeline along with the Tensors.

There’s a possibility that RDNA 3 will be delayed to the first part of H2 2022 due to the 5nm wafer supply from TSMC being tight, but looking at existing yields, that seems unlikely. This would make the Radeon RX 7000 GPUs the second product lineup after’s Apple’s M1 SoCs to leverage the highly advanced 5nm EUV process.

As for which GPUs to expect, the higher-end Navi 31/33 based RX 7800 XT/7900 XT are said to be launching first with the flagship offering packing 80 CUs similar to the RX 6900 XT. Since we’re looking at a die shrink, this means that either the die size will be smaller compared to Navi 21 or the additional die space will be dedicated to ray-tracing acceleration hardware. Either way, it appears that AMD will be launching its next-gen graphics cards quite a bit earlier than NVIDIA which is expected to delay the RTX 40 series (Ada) to Q4 2022 or possibly even Q1 2023.

Sources: 1, 2


Computer Engineering dropout (3 years), writer, journalist, and amateur poet. I started my first technology blog, Techquila while in college to address my hardware passion. Although largely successful, it was a classic example of too many people trying out multiple different. Left late 2019 and been working on Hardware Times ever since.
Back to top button