While it’s still hard to get an RTX 30-series or Radeon 6000 series GPU at MSRP in the market, though the condition is improving considerably, you cannot just ignore the rumors surrounding the upcoming GPUs from both AMD and NVIDIA.
There are a lot of expectations and anticipation regarding these new RDNA3 and RTX 40-series cards because this time the performance bump which almost everyone is murmuring is around 2x times what it is today. Yes, both NVIDIA’s ADA Lovelace and AMD’s Navi 30 series cards will adopt the multi-die GPU concept to increase performance substantially.
AMD has already done this on its server GPUs and NVIDIA is expected to do the same with upcoming Hopper architecture, so, it’s safe to say both GPU giants are on par with each other this time. The uplift in performance AMD showed last time with RDNA2 was way out-of-the-box for hardware enthusiasts like us and NVIDIA as well.
Having used all-new AMD GPUs and most of them from NVIDIA, the performance is really good and smooth if I compare with last-gen especially the ray tracing performance of RTX 20-series was awful. Also, both having their own upscaling technique has helped the gaming community to increase frame rates without sacrificing performance.
The rumors about the next-gen performance bump, especially for AMD started with around a 3-3.5 times performance increase compared to the current Navi 21 GPU. However, the tipster Greymon55 addresses the fact that if we keep a realistic expectation, then this upcoming Navi 31 flagship GPU will be 2.1 to 2.2x faster than the current Navi 21 GPU.
On the contrary, the upcoming flagship RTX 4090 is also expected to be 1.8-2x times faster than the current RTX 3090 which seems reasonable. So, on paper, AMD will match or could have a slight chance to exceed NVIDIA’s raw performance, but there are no words on the ray-tracing improvements till now. Please take these rumors with a pinch of salt but as time flows more accurate rumors could indeed bring some limelight to us.