A YouTube channel, Moore’s Law Is Dead (MLID), the upcoming RTX 4090 GPU with AD102 chip might be the first consumer GPU to come with GDDR7 memory. It will boost both performance and energy efficiency. Nvidia’s latest flagship GPU can have 18432 CUDA cores; it is the highest number of CUDA cores on any graphics card. It is said that RTX 4090 GPU will offer around 80 to 100 percent rasterization increase as compared to RTC 3090. The RTX 4090 will be two times faster than RTX 3090. It is rumored that this flagship will have an on-air TDP of around 450-600W. The same processor, AD102, will be used on RTX 4080. According to MLID, Nvidia has hugely improved the memory controllers for RTX 4000 and occurs to be designed in a way that they will support GDDR7 VRAM. The AD103 processor might be used on RTX 4070 and RTX 4060 Ti; it contains 10752 CUDA cores. It will have 256-bit GDDR memory. However, it is not known whether it will use GDDR7 or GDDR6x. While talking about performance, RTX 4070 is expected to be 20 to 30% faster than RTX 3090 (current),
Nvidia GeForce RTX 4060 (AD104)
RTX 4060 will have 7680 CUDA cores and 192-bit GDDR video memory. While talking about performance, it is powerful, just like RTX 3090 Ti. It is the most powerful graphics card at the time, and it is the same for ray-tracing capability.
Nvidia GeForce RTX 4050 Ti
It is expected to come with an AD106 chip and 4608 CUDA cores. This GPU is pushed to equal the performance of RTX 2080 Ti and RTX 2080 Super and will have 128-bit GDDR video memory.
