Four-slot heatsink rated to dissipate ~800 watts, four-slot design and PCB rotated 90°. This was supposed to be preparation for the GeForce RTX 4090 Ti, or the new Titan RTX, which would offer 10-20% higher performance than the current GeForce RTX 4090. Already after its release, there were reports that Nvidia CEO Jen-Hsun Huang was ready for all eventualities and wants to beat the competition at any cost (that is, even at the price of issuing a card with a default TDP of 600 watts+).
But the competitive offer did not turn out the way the competition (meaning AMD; Intel in the context of this performance generation probably does not need to be discussed) imagined. Chips from the series Navi 3x due to an unspecified bug, they cannot reach the planned frequencies in 3D load, and according to some sources, a patch of the bug, which caused graphic artifacts in some situations, took another few percent of their performance. Whatever the truth may be, in short, the performance turned out as we know it, and in the end it was decided that a refresh to fix the problem would not take place, because the withdrawal of engineers from the development of the next generation could negatively affect it as well.
GeForce RTX 4090 will probably remain unsurpassed in performance for two whole years until its release (Nvidia)
For Nvidia, this means that the GeForce RTX 4090 remains (and will remain until the release of the next generation) performance-free, so preparing a model with four slots and 600W consumption does not make sense. Especially in the current situation, when the overall demand for graphics cards is at a low level and the sales are dominated by the mainstream of the last generation (which today can be understood more or less as low-end), releasing a ~$2000 product does not make sense.
Reliable leaker kopite7kimi, according to which a new top model is no longer planned, but mentioned the preparations of two lower models. Nvidia is preparing a new variant of the GeForce RTX 4070 built on the AD103 GPU (this is currently used for the GeForce RTX 4080) and also a new variant of the GeForce RTX 4060 equipped with the AD106 core (this is what the GeForce RTX 4060 Ti now carries). It seems that Nvidia has already accumulated partially defective chips in their warehouses, which do not reach the configurations for which they were primarily intended, but can be used well for (a rung or two) lower models. After all, this was also the case with the previous generation. It can be expected that the parameters of the cards will not fundamentally differ from the existing ones, they will simply carry a higher, albeit more trimmed, GPU.
The current generation (GeForce RTX 4000) will be with us for another year and a half. Some older reports indicated the release of the new generation before the end of next year (2024), according to newer ones, it is more likely to be released only at the beginning of the next year (2025), since TSMC’s 3nm process has come with some delay and production debugging is a little slower than it first seemed. There is therefore a possibility that a competitive offer will come to the market a little earlier, even on the 4nm process (but a lot can still change in a year and a half).
2023-07-28 05:48:46
#GeForce #RTX #Canceled #RTX #Versions #Coming
