Micron has announced production of its 16 gigabit (Gb) GDDR6X memory chips and claims to be the only memory maker to now ship the GDDR6X chips in large quantities. This is certainly remarkable as the capacity will be doubled. In addition, Micron claims that this new 16 gigabit (Gb) GDDR6X memory chip will provide 15% better performance compared to the previous 8 GB version.
These memory chips are the same ones NVIDIA has used on the newly released GeForce RTX 3090 Ti, the latest consumer flagship of the Ampere family. The big difference between the GeForce RTX 3090 Ti and its non-TI counterpart is the improved memory. Both models have 24 GB GDDR6X memory spread over a 384-bit bus, but what sets the Ti model apart is the use of the 21 Gbps memory chips. The non-Ti model, on the other hand, uses 19.5 Gbps. This represents a memory bandwidth of 1 TB/s as opposed to 936.2 GB/s on the latest flagship.
There’s no question that the GeForce RTX 3090 Ti is the fastest consumer graphics card NVIDIA has ever built. Thanks to the improved memory without which this would not have been possible. In addition to the improved memory, it also features more CUDA cores and a higher TGP, which contributes to better gaming performance. Interestingly, there are rumors that the GeForce RTX 3090 Ti with its 12-pin PCIe connector could likely be a test vehicle for NVIDIA’s next-generation GeForce RTX 40 series based on Ada Lovelace. If this is any truth, it’s not wrong to say that this is just a test run for future GDDR6X memory chips that will be increasingly faster and built with more capacity.