One of NVIDIA’s secret weapons when it came to the performance of its current graphics cards was the adoption of a proprietary Micron standard for VRAM. Well, it seems that the GDDR6X will repeat in the RTX 40. This time much faster and with greater capacity. Since the second generation of this memory has just been officially announced. What will bring us again?
Video memory or VRAM is crucial for the performance of a graphics card, which is why NVIDIA opted for GDDR6X for its high-end RTX 30, due to the fact that they can provide much higher bandwidth than GDDR6 conventional. Well, it seems that we can consider Micron’s VRAM for exclusive use to accompany NVIDIA’s most advanced GPUs as a success.
And it is that the company specialized in the design and manufacture of memories has just announced the second generation of said memory. Let’s see, therefore, how the GDDR6X will be in the NVIDIA RTX 40 and what new performance and capacity it brings us compared to the current one. Will we see AMD finally adopt this type of memory in its future RX 7000 or will it continue to be exclusive to the green logo brand? Who knows, for the moment we bet to see it only in the next generation of the green brand, but we could be in for a surprise.
This is the GDDR6X for RTX 40: 2 GB per chip and 24 Gbps speed
If there is something that has been highly criticized about the GDDR6X during the NVIDIA RTX 30 generation, it is the fact that its capacity per memory chip has been 1 GB for a long time. So it was necessary to place twice as many chips, in shared data bus mode, in order to have very high memory capacities. Which doubles the cost of memory on the graphics card. However, that GDDR6X exists with 2GB capacity per chip is not a novelty, as is the case with the RTX 3090 Ti.
It is the increase in bandwidth where the GDDR6X takes all the force, since when it becomes 24Gbps We are talking about today’s hypothetical RTX 3090 with a 384-bit bus that would go from having a bandwidth of 1008 GB/s to having a bandwidth of 1152GB/s. In any case, we keep wondering if this will be enough to feed the up to 144 shader cores that this beast of a graphics card will have.
Nor can we forget that this improvement has been made possible by the increase in power consumed, thanks to the new 12+4-pin connector. Thanks to which the graphics cards will no longer have a consumption limit of 375 W at 675 W. At the same time, the question that assails us is to know if for the mid-range of the RTX 40 we will also see the use of the GDDR6X this time , although we expect it at transfer speeds lower than that 24 Gbps.