Nvidia will launch a new GeForce RTX 3080 in January next year, with 12GB of video memory

by ·

Recently, it has been circulated that NVIDIA will launch GeForce RTX 3090 Ti (GA102-350), GeForce RTX 3070 Ti 16GB version (GA104-401), and GeForce RTX 2060 12GB version (TU106-300) in January next year. On the desktop platform, the appearance of these products has put a big question mark on the so-called GeForce RTX 30 SUPER series, but it seems to be more than that.

MSI GeForce RTX 3080 Ti SUPRIM X

MSI GeForce RTX 3080 Ti SUPRIM X, Source: VideoCardz

According to VideoCardz reports, Nvidia has notified its partners that in addition to the previous graphics cards, a new version of GeForce RTX 3080 will be launched at the same time, using GA102-220 cores, and the memory capacity will be increased to 12GB. The earliest GeForce RTX 3080 used a GA102-200 core, with 8704 stream processors, equipped with 10GB of GDDR6X video memory, at a rate of 19 Gbps, and then added the introduction of a new version of the cryptocurrency mining limiter, which became GA102- 202 core. However, this is not the first time that the GA102-220 core has appeared on the Internet. Twitter user @kopite7kimi has mentioned it before.

Due to the change to 12GB of GDDR6X video memory, it means that the video memory bit width has also become 384 bits. The new version of GeForce RTX 3080 will increase both the video memory capacity and the bandwidth, but NVIDIA has not confirmed the video memory rate with partners. In addition, the number of stream processors in the GA102-220 core configuration is temporarily uncertain, and it is not clear whether to maintain 8704 or to increase. In addition, it is rumored that the GeForce RTX 2060 12GB version will use the PCB design of the GeForce RTX 2060 SUPER.

If the above news is true, then the entire GeForce RTX 30 (LHR) series product line of Nvidia’s Ampere architecture will have six models involving GA104 and GA102 cores on consumer-grade desktop platforms. On these new GPUs, NVIDIA seems to be more concerned about the increase in memory capacity or speed, and optimize the GPU memory configuration.

Tags: