The NVIDIA GeForce GTX 10 Series graphics cards based on the Pascal architecture were a huge leap in performance compared to their Maxwell-based GeForce 9 Series predecessors; 16nm chips provided a significant improvement in terms of performance, efficiency and overall value, also marking one of the biggest advancements in ‘Ti’ graphics performance. To this day, the GTX 1080 Ti it is considered the graphics card ever created that represented a performance leap that NVIDIA has not been able to match with its Turing and Ampere flagships.
The biggest generational leap ever seen, with the NVIDIA RTX 40 Series
The NVIDIA GeForce RTX 40 Series is expected to deliver the same generational performance leap as the GeForce RTX 30 series; Based on the new Ada Lovelace GPU architecture, this next generation graphics cards will use the node at 5 nm from TSMC And while they will consume quite a bit of power in comparison, their efficiency ratings will also be vastly increased, which will be an unparalleled performance leap.
There are rumors that mention some other details regarding the operating speeds and the supposed release date. We know that the RTX 40 series is a long way off at the moment, and the latest rumor claims that we do not expect these charts to arrive in Q4 2022 as expected but rather that they will arrive in early 2023.
In terms of operating speeds, NVIDIA GeForce RTX 40 Series GPUs with Ada Lovelace architecture are said to offer speeds between 2.2 and 2.5 GHz in Boost mode, which would be a nice improvement over Ampere’s 1.7-1.9 GHz. Pascal also saw an impressive increase in speed and was in fact the first GPU architecture to break the 2 GHz barrier (and yet today it is AMD that has the ‘throne’ of speed in its RDNA 2 GPUs, which reach high speeds. over 2.5 GHz).
What we know about Ada Lovelace
But do not think that 2022 will be a dry year for NVIDIA, as this is also due to the fact that NVIDIA will reportedly offer an update SUPER middle of its current line of RTX 30 Series over the next year. Therefore, if that line is presented before the RTX 40 Series (which would be logical), then we can expect a delay in its launch in order not to “step” on the launch of these SUPER variants. This, at the same time, means that AMD could have its line of graphics cards ready. RDNA 3 by the time NVIDIA launches its next generation of GPUs.
Based on these numbers we mentioned earlier, if the RTX 4090 or whatever comes out with the AD102 GPU features 18,432 CUDA cores, we can wait until 80 TFLOPs FP32 compute throughput at 2.2 GHz (and more than doubled in FP16), which is a sheer outrage that would leave the current RTX 3090 at the height of the bitumen. These numbers align with rumors that we can expect a performance jump of up to 2.5 times with these future RTX 40 Series.
|Node||TSMC 12nm NFF||Samsung 8nm||TSMC 5nm|
|Graphics Processing Clusters (GPC)||6||7||12|
|Texture Processing Clusters (TPC)||36||42||72|
|Streaming Multiprocessors (SM)||72||84||144|
|Flagship||RTX 2080 Ti||RTX 3090||RTX 4090?|
|Launch||March 2017||September 2019||2022-2023 (To be confirmed)|
All in all, the most important question in this regard remains whether the RTX 40 Series will also carry the same under the arm. price that we saw with the Pascal architecture. Not only was the GTX 1080 Ti the best graphics card for performance, it also offered the best value of all NVIDIA “Ti” GPUs, which is why Pascal is considered one of the best generations in NVIDIA history.
The $ 699 introductory price for the GTX 1080 Ti made this GPU incredible value, but NVIDIA also lowered the price of its standard line to $ 499 for the GTX 1080 and $ 349 for the GTX 1070, and definitely with the price of the current graphics we would love to see that NVIDIA returns to this trend, much more competitive and that it benefits us users. However, with the increased cost of components and logistics, the price of Ada Lovelace GPUs is likely to be the opposite: more expensive than the current RTX 30 Series.