The older version of GeForce RTX 4080 was 24-30% faster than the younger version — NVIDIA tests

The older version of GeForce RTX 4080 was 24-30% faster than the younger version — NVIDIA tests

NVIDIA shared the new results of the video cards GeForce RTX 408016GB and GeForce RTX 408012GB conducted by the company ' s specialists. The new game performance was tested in three games that were launched in the 4K resolution.

====Requiem, F122 and Microsoft Flight Simulator=====Speeds were tested in A Plague Tale games: Requiem, F122 and Microsoft Flight Simulator. The games were launched with or without the DLSS scale technology. According to the tests, the GeForce RTX 4080 with 16 Gbyte memory was faster than the 12 Gbytes of memory by 24 to 30 per cent. This significant difference in speed can be attributed not so much to the availability of more memory in one model as to the use of different graphic processors in the video cards.

The GeForce RTX 408016GB uses an AD103 graphic processor with 9728 CUDA kernels. In turn, the GeForce RTX 408012GB model uses an AD104 graphic chip with 7,680 CUDA kernels. This is 26% less than the 16 Gbyte memory.

There is also a significant impact on the performance of the memory tyre. For a map with a large amount of memory, it is 256 bits and for the lower version 192 bits. In addition, the accelerators are different in terms of energy consumption. For the older model, 320 W and 285 W are declared.

It appears that GeForce RTX 408012GB is not much higher than GeForce RTX 3090 Ti, at least without the use of DLSS 3. At the same time, the new 10-20% faster than the GeForce RTX 3080 model.

GeForce RTX 408016GB and GeForce RTX 408012GB will be released in November.