Although the next-gen GPUs will arrive on the market only at the end of the year, AMD and NVIDIA begin to challenge each other on the “rumor” side, leaking some preliminary data that, in the context of the many rumors circulated in recent weeks, suggests a comparison without no holds barred. Nothing new we could say, however the latest news on the net from those who are essentially the most trusted leakers in the sector open the door to an unprecedented generation of consumer graphics cards, able to almost ridicule the current products on the market.
This is what emerges from the recent tweets of kopite7kimi and Greymon55 regarding what will be the top of the range proposals of the Radeon RX 7000 “RDNA3” and NVIDIA GeForce RTX 4000 “Ada Lovelace” series; according to what has been reported, the performance bar will rise so much that the two big names in the sector will come to propose GPUs capable of guaranteeing computing power in the order of 100 TFLOPs (FP32 calculation).
No reference regarding models and features, but we know well that we are talking about the alleged GeForce RTX 4090 and Radeon RX 7900 XT . Remembering that we are dealing with unofficial or definitive data, this would mean a 2.5 times increase in performance compared to the very powerful GeForce RTX 3090 Ti , while if we look at the AMD counterpart it would quadruple the computing power of the Radeon RX 6900 XT .
To be honest, I don't have much information about AMD. Maybe Lisa and Jensen's competition will give us a 100TFLOPS gaming war in a few months.
— kopite7kimi (@kopite7kimi) April 29, 2022
In addition to the implementation of new architectures on more advanced production nodes (5nm), it is rumored that next-gen GPUs will have very high frequencies: GeForce RTX 4090 (GPU AD102) is credited with a clock at 2.7 GHz , Radeon RX 7900XT (Navi 31 GPU) instead could even touch 3 GHz . Values to say the least extreme given the nature and technical characteristics of these chips which, consequently, will inevitably affect consumption; for the GeForce RTX 4000 there has been talk of a maximum TBP of 600 watts several times , while on the Radeon RX 7000 side there are few data which, however, do not exclude that AMD can go in the same direction (however, let’s think in a less extreme way).
It is still too early to make further considerations, but after about a semester the ingredients seem to be all there for a challenge that we expect to be very heated. The performance factor will be decisive, but from the users’ point of view, peak consumption could have a considerable weight, at least as regards the enthusiast range.