Nvidia’s GPU based on the Ada Lovelace architecture may be one of the most important releases in 2022, which corresponds to the GeForce RTX 40 series, and it is rumored that a new game graphics card for players will be released in the fourth quarter of 2022.
According to Twitter user @greymon55, the performance of the Ada Lovelace architecture will be doubled, but with that, the power consumption will also be doubled. Earlier news said that the Ada Lovelace architecture GPU will be manufactured using TSMC’s 5nm process, and the flagship product AD102 is a 600mm² large chip. This GPU will be equipped with 18432 CUDA cores, the performance is doubled on the basis of GA102, with GDDR6X video memory, the bit width is 384 bits. It is rumored to have a core frequency of 2.2 GHz and provides 81 TFLOP of computing performance (FP32). The competitor of the GeForce RTX 40 series will be AMD’s Radeon RX 7000 series, which is also expected to be launched at the end of 2022.
Unlike AMD, Nvidia’s Ada Lovelace architecture GPU will retain the traditional design and will not use MCM multi-chip packaging like AMD’s RDNA 3 architecture-based Navi 31. AMD’s RDNA 3 architecture-based Navi 31 is AMD’s first consumer GPU with this technology. Nvidia temporarily only uses MCM multi-chip packaging on the Hopper architecture GPU, which is different from the Ada Lovelace architecture GPU. This is a product for data centers and AI computing.
It is rumored that GeForce RTX 3090 Ti will use a GA102-350 core with a TDP of 450W. It will be the first product to support the new PCIe 5.0 external power supply interface. It is expected that the flagship GeForce RTX 40 series graphics cards based on the Ada Lovelace architecture GPU will have a TDP of 450W or above, or even 600W. This also explains why NVIDIA is so anxious to promote the new PCIe 5.0 external power supply interface. The performance improvement of the GeForce RTX 40 series this time is likely to come at the cost of increasing overall power consumption.