Nvidia RTX 4000 series increases cache by 16x

The hacker group that attacked Nvidia’s servers has clarified many issues that the game world is curious about. After the core numbers of the Nvidia Ada Lovelace architecture, the cache capacities were also announced.

Big leap in Nvidia Ada Lovelace architecture

We know that Nvidia Ada Lovelace architecture will offer up to two times performance increase compared to Ampere architecture and power consumption up to 850W. The leaked new documents also revealed the cache numbers after the number of cores.

The new graphic to be named AD100 unit does not differ from the previous generation in terms of bus, but the AD104 core is downgraded to 192-bit, while the AD106 core is downgraded to a 128-bit bus.

There is an explosion on the memory side. As the top model for gamers, the AD102 core has 96MB of Level 2 cache and we see an increase of around 16 times. Other cores also see huge memory gains.

Nvidia, which wants to experience an assertive rise in the new generation with faster graphics memories, feels the need to make a move against AMD’s steps. The fact that the new generation Radeon graphics cards will use 256MB-512MB of Infinity memory will further intensify the competition.