There are some fresh reports circulating about the essential features of the graphics cards that are supposed to lead the Ada Lovelace graphics architecture. The most recent speculations concerning the GeForce RTX 4090, 4080, and 4070 have been shared by Kopite7kimi from their sources. The RTX 4090, which is now said to have base/boost clocks of 2,235 / 2,520 MHz, is the focus of the most important new information.
If accurate, the default GPU clock speeds of the Nvidia GeForce RTX 4090 are surprisingly high. The GeForce RTX 3090, the card it replaces, had substantially slower GPU clocks (1,395/1,695 MHz for the reference architecture). The new Ada Lovelace flagship will also have quicker memory (the same 24GB limit) and approximately 6,000 more CUDA cores, but if the rumored TDP is accurate, it will consume about 100W more power.
It should be emphasized that, if our calculations are accurate, the RTX 4090’s computational performance maybe 90 TFLOPs or such. Comparing that to the RTX 3090’s 36 TFLOPs is a huge improvement. Kopite7kimi further claims that the AD102 GPU can be cranked up to 2,750 MHz.
A 450W card, the Nvidia GeForce RTX 3090 Ti, is purportedly the same as the RTX 4090.
You can see the variety of cooler sizes and types in the current crop of RTX 3090 Ti cards, so the cooling solutions won’t be any different from those we’ve seen before.
It remains to be seen whether Kopite7kimi’s data is accurate. The specifications for the RTX 4090 were slightly different in a report we released based on his data in mid-May, but we always advise adding some salt to these leak claims.
It’s also possible that the actual retail GPU specifications are still being determined by Nvidia and its partners, who are still deciding on the enabled cores and clocks for delivering products.
Also Read:
Virgin Orbit successfully launched its Rocket Carrying Seven Satellites
