The newly emerged data from the Kopite7kimi insider indicate that the NVIDIA GeForce RTX 4090 has a base frequency of 2235 MHz and an acceleration of 2520 MHz with an actual clock frequency of 2750 MHz.
These characteristics are impressive compared to the existing GeForce RTX 3090. The base frequency has increased by 60%compared to 1395 MHz, and the acceleration frequency increased by 49%compared to 1695 MHz. This is really a real improvement compared to a series of video cards on AMPERE architecture.
The GeForce RTX 4090, according to rumors, has a configuration of CUDA 16384 nuclei, this means that with the indicated "actual" clock frequency of 2750 MHz, the computing performance of the card can reach 90 TFLOPS during working load with single accuracy. This is almost exactly two times higher than that of the GeForce RTX 3090.
The source also speaks of the past specifications of the GeForce RTX 4080 and RTX 4070 - each model should have its own graphic processor, it is either AD103 (RTX 4080) or AD104 (RTX 4070). It seems that the RTX 4080 will be equipped with a 256-bit memory bus and 16 GB of GDDR6X memory with a clock frequency equal to the speed RTX 4090. Meanwhile, RTX 4070, which, according to the latest rumors, presumably has 7168 cuda cores, will have a 160-bit tire Memory and 10 GB of memory GDDR6 18 Gbit/s.
Most likely, NVIDIA will present its series of GeForce RTX 4000 around September or October of this year.
On this I have everything, good luck to everyone!
Hold the author with a penny 🙃 | 4377 7237 8152 3315
🔴 If you were useful to sign this article!
🔴 Are there any questions? Leave the comment and I will answer it!
🔴 Did you like the article? Click your finger up!
#NVIDIARTX #AMD #intel #PC #video card #processors #assembly