To start the week off right, what might be better than some new rumours/information regarding NVIDIA’s next flagship graphics card: the RTX 4090. According to this latest rumor from leaker greymon55 on Twitter, the RTX 4090 would have 100 TFLOPS of FP32 (single precision) power. Thus, the AD102 GPU composing this next card might exceed the threshold of 100 TFLOPS. In comparison, the RTX 3090 Ti yet very powerful, is only 40 TFLOPS of power in FP32. On the other hand, be careful, it must be understood that this measure is not proportional to the gains that we will have in gaming use.
NVIDIA GeForce RTX 4090: 2.5 times more raw computing power than the RTX 3090 Ti?
To achieve such power with its 18,432 CUDA cores, the AD102 GPU should operate at a frequency of 2.7 GHz. But with a partially disabled GPU on the RTX 4090, the frequency should be even higher. Note that in parallel, the latest rumors regarding the graphics card AMD Radeon RX 7900 XT (Navi 31) mention 92 TFLOPS of FP32 power for a frequency of 3 GHz. NVIDIA and AMD cards would therefore potentially be in the same frequency ranges.
Remember that regarding the release of these cards, whether AMD or NVIDIA, there is talk of a Q3 or Q4 launch of this year. The competition will therefore be tough and we can’t wait to see which one will be the most successful. Indeed, there are still too many unknowns at this time to make good estimates and some of the information we currently have may well still change between now and the launch of the GPUs.
I can only say that the two products have improved a lot compared to their predecessors, but if you want to ask me directly which one is better, I’m sorry I can’t answer, because no one knows the specific improvement by percentage.
— Greymon55 (@greymon55) April 30, 2022
But let’s also remember that TFLOPS are not everything. We must also look at the technologies around such as Ray Tracing or technologies such as DLSS for NVIDIA or FSR for AMD.
