GeForce RTX 40 “Ada” GPUs Will Feature Large L2 Caches

Leaked data by the hackers provides substantial information of NVIDIA’s next architecture codenamed “Ada”. On the other hand, NVIDIA is still struggling to figure out how the hackers obtained so much information.

NVIDIA RTX 40 GPUs are expected to possess 96 MB of L2 Cache, up to 90 MBs more than Ampere RTX GPUs. The leaked files revealed this latest information last week. The data supposedly verifies that “Ada” GPUs will be equipped with 16MB of cache per 64-bit memory bus, considerably better than Ampere with 512 KB per 32-bit memory.

NVIDIA AD102 GPU Diagram

Moreover, the leak revealed that the flagship Ada GPU codenamed AD102 is to possess 96 MB of cache, AD103 and AD104 will have 64 MB each, while AD106 and AD107 would be equipped with 48 MB and 32 MB respectively.

This latest information from the leaks along with information from credible NVIDIA leaker Kopite7kimi, have been sufficient for this proposed GPU block diagram to be confirmed for the upcoming AD102 GPU. According to reports, the chip would have a die size of 600 mm², almost the same size as GA102.

RTX3050 Hero Banner Big 1200x344 1

The RTX 40 GPUs is expected to get 71% more cores than RTX 30 GPUs, that too when each chip will see its full implementation released to the market. Features like clock speeds and power requirements for new cards are not clear yet, but it is quite obvious that RTX 40 GPUs will be a considerable upgrade over Ampere. The same architectural upgrade is anticipated from AMD RDNA3 GPUs.

Via VideoCardz