According to various rumors, the upcoming Nvidia GeForce RTX 2060 will soon be reissued with 12 GB of VRAM. The refresh model will differ from the original not only in terms of the larger graphics memory.
After there were already many rumors about a possible refresh of the GeForce RTX 2060 with 12 GB VRAM, more information about the graphics card has now been published by the Polish tech blog Purepc.pl. It turned out that the upcoming GPU differs from the original not only in terms of the larger graphics memory.
According to this, the new Nvidia GeForce RTX 2060 should have 2176 CUDA cores, while the original model managed with 1920 cores. This means that the RTX 2060 with 12 GB of VRAM would have as many CUDA cores as the RTX 2060 SUPER. The boost clock will also be 1650 MHz and the memory will be connected with a 192-bit interface. At the same time, the TDP should be slightly higher at 184 watts, which is interesting because none of the changes at first glance justifies a higher TDP than the RTX 2060 SUPER with 175 watts.
If these specifications come true, the upcoming version should be just below the GeForce RTX 3060 in terms of performance. It will be exciting to see whether Nvidia can really do something about the chip shortage with this step. Either way, this should be financially worthwhile for the manufacturer, as the demand for GPUs currently significantly exceeds the supply and even old models are therefore very popular.
A mix of RTX 2060 and RTX 2060 Super
With reference to three unspecified sources, the website claims that the RTX 2060 with 12 GB should not only double the memory but also increase the number of shaders. Accordingly, the RTX 2060 should appear 12 GB with 2,176 shaders and thus offer just as many computing units as the RTX 2060 Super. In order to be able to control the 12 GB, the 192-bit memory interface of the regular RTX 2060 remains. The supermodel, on the other hand, comes with 256 bits.
| graphic card | GPU | Shader | Boost clock | TDP | interface | Storage |
|---|---|---|---|---|---|---|
| RTX 2060 | TU106 | 1,920 | 1,680 MHz | 160 W | 192 bits | 6 GB GDDR6 |
| RTX 2060 Super | TU106 | 2.176 | 1,650 MHz | 175 W | 256 bit | 8 GB GDDR6 |
| RTX 2060 12 GB | TU106 | 2.176 | 1,650 MHz | 184 W | 192 bits | 12 GB GDDR6 |
According to Videocardz, the clock rates of the RTX 2060 12 GB should also correspond to those of the RTX 2060 Super. In the boost, the built-in GPU should run at a maximum of 1,650 MHz, overclocked models are supposedly planned with up to 1,690 MHz. Compared to the supermodel, there is also a slightly larger TDP range of 184 instead of 175 watts. Unfortunately, it is not yet entirely clear why the higher TDP is needed. Nvidia may install 12 GB of faster and therefore more hungry GDDR6 memory on the RTX 2060 than on the other models in order to compensate for the memory interface, which is a bit narrow compared to the performance.
Whether the key data presented here actually apply in the end will probably only be clarified when the graphics card is officially launched. This is supposedly planned for December 7th and thus Tuesday in two weeks. Just in time for the Christmas business, there should be a new alternative to the RTX 3060, which has the same amount of memory but at the same time has a little less performance and older architecture.
