To begin to understand these graphics cards in greater detail, we must understand that they provide their GPUs first, which, as we well know, is the heart of each.
After all, all the other parameters such as VRAM or their speed and power will be the only evidence compared to the power they are facing, but at the same time they usually decide to make a difference, a really complex relationship at this distance.
RX 5700 vs RTX 2060 SUPER: battle going up the ladder
Of these two graphics cards we get the best of both AMD and NVIDIA in terms of properties, because in the first case you get a new one RDNA 1.0, and in the case of secondary storage Exercise The primary asset.
Where we can't find the similarities is in their lithographic processes, because the AMD card gets a new far more recent location 7 nm produced by TSMC and life-saving for its Navi 10 Pro GPU, while on the NVIDIA card, TSMC is also responsible for providing the TU106-410-A1 chip, but 12 nm, so AMD is well defined in this section.
The lithographic process has an effect on the size and transistors of each GPU. Therefore, the RX 5700 reaches a smaller size 251 mm2 in front of 445 mm2 of the RTX 2060 SUPER is approximately the same number of transistors: 10.3 billion for AMD and 10.8 billion for NVIDIA.
This leaves us the same amount of another size for 77.29% among them, it underscores the importance of the lithographic connection between them.
High fit in your specification
Starting with the specification section itself, we get a 36 CU RX 5700 computation, where its RTX 2060 SUPER competitor in its segment reaches 34 SM.
The number of lower units yields the total 2304 shaders vs 2176 shaders, so the AMD card handles placement on paper with 5.88% more calls. The TMU figure also likes the Lisa Su card compared to Huang, due to the large number of shaders and the amount to be followed by these two parts.
Therefore, the RX 5700 is rated at 144 TMUs compared to the 136 TMUs of the RTX 2060 SUPER. Surprisingly, there is a perfect fit for their ROPs, since they both carry 64. This is due to certain limitations that RDNA still draws on its previous building materials, and that RDNA 2 should eliminate this.
As for RT Cores and Tensor Cores as units of AI and Ray TrackingAMD does not have this architecture, but NVIDIA reaches 34 and 272 units respectively, so the red card cannot perform any of the two functions used to accelerate DeepLearning or ray tracing.
One is to know that they share the same L2 c2 size, 4096 KB, so the assignments and exchange of information between GPU, VRAM and PCIe will have the same power on all chips.
The frequencies will decide.
As for the speed carried by both GPUs, it is necessary to influence the fact that AMD has included a single frequency on their cards. So, now it has come Basic, Game and Boost, while only NVIDIA has the first and last.
The curious thing about this new AMD speed is that its cards, in effect, fail to keep any of the three named waves, usually in the distance between the game and Boost.
NVIDIA, on the other hand, achieves the opposite, always achieving higher frequency than that announced in Boost. This is important to consider when looking specifically at these frequencies, since the RX 5700 reaches 1465 MHz, 1625 MHz in the Game and 1725 MHz in Boost, while the RTX 2060 SUPER reaches 1470 MHz and 1650 MHz respectively.
In any case, both cards use the same GDDR6 power (8 GB) and speed (14 Gbps or 1750 MHz real). They also share a data bus (256 bits) so the result is that their bandwidth is exactly the same, 448 GB / s.
The difference starts to finally shine
Although as we've seen they share many of their features, the fact is that both cards are at different stages when we consider their FP32 performance. The AMD RX 5700 reaches 7.95 TFLOPS, while the RTX 2060 SUPER reaches 7.2 TFLOPS
This functionality has no purpose of specifying the different types of PCIe supported by both cards, because the AMD GPU is compatible PCIe 4.0 16X, while NVIDIA has access to PCIe 3.0 16X. And if in fact it was a deciding factor, the double performance achieved with version 4.0 compared to 3.0 would leave an entirely different TFLOPS.
Where we will find the difference is quite impressive in terms of power consumption and connectivity, since the RX 5700 even though it has the most advanced node achieves a TDP of 180 watts with an 8 + 6 pin system.
The RTX 2060 SUPER is surprisingly good, since it reaches TDP for 175 Watts with only 8 connectors.
In the end, we have pricing, a deciding factor in any current graphics card due to the rise of both companies. The AMD card is provided from $ 349, when that NVIDIA was introduced in price $ 429, which has now gone down into the illegal zone between $ 375 and $ 400 On average
Working with value
The differences between both cards are quite small in terms of performance, but the reality is that currently in global terms, the NVIDIA option is one step further than AMD in most games.
As for the scale performance / dollar, after the NVIDIA option price changes, this battle is well illustrated, where it depends on the model we choose to go down one or the other.
Therefore, the choice of card is important to be able to give the winner the most obvious because of all the specifications, which will also depend on the preferences or most importantly the shouting or temperature of each user and equipment.
Table of Contents
