AMD has finally talked about its next generation "Big Navi" graphics cards, which from now on in will be called RDNA 2 or "Navi 2X" -- I kinda liked the name Big Navi, though. AMD CEO Dr. Lisa Su did say that AMD was preparing Big Navi to compete with NVIDIA, as the high-end GPU market is "very important" to the company.
During its recent Financial Analyst Day, AMD came out swinging talking about its warpath against its competitors in Intel and NVIDIA -- where it revealed that its next-gen discrete graphics cards will be based on the RDNA 2 architecture.
The company is referring to RDNA 2 as "Navi 2X" which will arrive on the 7nm node, but it also teased RDNA 3 or "Navi 3X" on an "Advanced Node", too. RDNA 3 / Navi 3X should be on the 7nm+ node, or the 5nm node -- something I'm going to need to clarify.
The new RDNA 2 architecture will have hardware-based ray tracing features, just like NVIDIA's current-gen Turing GPU architecture. RDNA 2 is part of the custom silicon AMD is providing for the SoCs going into Microsoft's next-gen Xbox Series X, and Sony's PlayStation 5 consoles. AMD said that RDNA 2 / Navi 2X will give gamers "enthusiast-class" performance, kick ass power efficiency, and a bunch of high-end GPUs that will offer "uncompromising 4K gaming".
RDNA 2-based graphics cards will not only support hardware-based ray tracing, but they'll also be capable of Variable Rate Shading (VRS). This is something Microsoft itself confirmed during its Xbox Series X hardware tease.
We should expect the flagship RDNA 2-based graphics card to be the rumored Radeon RX 5950 XT, which according to the last rumors packed a huge 80 compute units, 5120 stream processors, and up to 17.5 TFLOPs of compute performance. If that happens, I really want to see AMD push forward with the Radeon RX 6000 series over the Radeon RX 5000 series which is powered by the current-gen RDNA 1 architecture.
AMD RDNA 2 / Navi 2X / Big Navi Specs
- Optimized 7nm+ node (kinda confirmed)
- RDNA 2 architecture (confirmed)
- 80 compute units
- 5120 stream processors
- 17.5 TFLOPs compute performance
- Hardware ray tracing support (confirmed)
NVIDIA is currently cooking its next-gen Ampere GPU architecture, something that will power the GeForce RTX 3080 and GeForce RTX 3070 that were recently teased. These cards have a huge 10/20GB of GDDR6 on the higher-end SKUs, while rumor has it the lower-end cards will have 8/16GB GDDR6.
Check out these monster specs:
GA103 (GeForce RTX 3080)
- 10/20GB GDDR6
- 320-bit memory interface
- 60 SMs
- 3480 CUDA cores
GA104 (GeForce RTX 3070)
- 8/16GB GDDR6
- 256-bit memory interface
- 48 SMs
- 3072 CUDA cores