The brand new RTX 3090 at the time of writing this article, the most powerful graphics card NVIDIA has on the market within the GeForce range; however, there are indications that when the time comes it may appear a more powerful graphics card.
The RTX 3090 Ti will use the full suspension of the GA102 chip
The core chip of the RTX 3090 is a beast called GA102-300 with a surface area of 628mm.2 and 84 SMs in total.
What if we tell you that the RTX 3090 does not use a full 84 SM configuration but one with 82 SM? We are actually talking about performance of just over 2.5% at the same speed in all aspects, but there are still 2 disabled SMs that can add performance enhancement.
We must keep in mind, however, that as cooling production will be easier to get graphics cards with a GPU with a full and stable setting at a much higher speed than those of the current RTX 3090. that currently only 10% of the manufactured drawings are used to build Founder Edition models, but over time, this percentage will increase as the production area is polished.
The RTX 3090 Ti will use the faster GDDR6X
In memory, graphics cards are not affected by latency but are vulnerable to the bandwidth they receive. The reason for this is that in processor processors, or processor throughput, their performance is related to bandwidth, and if this is not large enough when the output of the GPU is not 100%. The effect of faster memory usage was seen with the GeForce RTX 20 × 0 Super range which had faster GDDR6.
NVIDIA could begin launching a new system next year making the most of GDDR6X’s fastest and highest quality memory. But instead it should be expected in 2021 with the launch of the latest RTX 3000 with fast GDDR6X memory because the standard supports up to 21 Gbps speed at the moment. Currently GDDR6X chips are integrated at 19 Gbps and 19.5 Gbps.
21 Gbps GDDR6X chips with very low pass rate; This means that for chips from each device, those that can support this definition are too small at the moment to make a memory card that memory speed.
Another problem is that the GDDR6X is currently sold in bulk of 1GB per chip where the 2GB GDDR6 is already available. This makes the GDDR6X memory more expensive and makes it necessary for the convergers to currently decide between high GDDR6 power and GDDR6X high speed, something that should be resolved with the introduction of GDDR6X chips with a maximum of 2 GB.
Improving the use of the GDDR6X is essential to the final adoption of GPUs range RTX 3000 still working with GDDR6, and not only can we expect the RTX 3090 Ti and 3080 Ti with GDDR6X memory at high speed but ultimately the RTX 3060 Ti and RTX 3070 Ti replacing GDDR6 controller with one GDDR6X.
Are You Super?
We do not know whether the upgraded RTX 3000 models will come out with the Ti or Super surname, as one thing or another may happen. We do not believe that NVIDIA will introduce the updated version of the RTX 3000 before launching all the representatives of that list.