We have been talking about RDNA 2 for a long time, a new generation architecture with which AMD does not aim to compete with the 20 series RTX, but rather will be designed to cope with RTX 30 series NVIDIA, whose launch is scheduled for sometime in September this year.
When AMD announced the RDNA architecture, it gave us both good and bad news. The good news was that they had started the path that will finally lead them to abandon the foundations of the GCN architecture, a true veteran who has been with us since 2011, when the Radeon HD 7000 arrived, and who had already shown clear signs of exhaustion.
I am convinced that the end of the GCN architecture should have occurred with the Radeon R9 Fury and Radeon R9 Fury X. These were the last two graphics cards that could really compete with the most powerful NVIDIA within that generation (the GTX 980 Ti, for more details), but AMD had nothing with which to displace said architecture, so it had no more Remedy to try to polish it to the extreme, a strategy that resulted in the Radeon VII, the last top of the GCN-based general consumer range.
The RDNA architecture is a hybrid that maintains a common base with the GCN architecture, something that AMD said was fundamental for optimization reasons, but represents an important step forward in terms of efficiency and performance. Just think, for example, that an RX 5700 XT with 2,560 shaders renders only slightly less than a Radeon VII with 3,840 shaders.
RDNA 2 will be the final goodbye to the GCN architecture
There is no doubt that it is about time. This architecture will have a completely new base and will mark an important turning point, not only because it will be present in the new generation consoles (PS5 will use an intermediate version, RDNA 1.5), but because it will allow AMD compete again with NVIDIA at all levels of the graphic sector of general consumption, and promises to have the latest advances in the sector.
In case any of our readers have been lost with this statement, I remind you that the most powerful AMD today, the RX 5700 XT, offers performance slightly higher than that of an RTX 2070But it’s no match for the RTX 2070 Super, which means it can’t compete at the top with NVIDIA.
That landscape will completely change with the arrival of RDNA 2, an architecture that, by completely breaking the foundations of the GCN architecture, will allow AMD develop graphics cards without the limitations imposed by that in the geometry and shading engines, which means, in summary, we will see Radeon graphics cards with more than 4,096 shaders and with more than 64 raster units. Good news, no doubt.
In addition to the corresponding improvement at the level of shaders and other elements, the RDNA 2 architecture is expected to bring a considerable increase in terms of CPI, and that it also offers a notable jump in terms of efficiency energy. In this sense, the use of a more mature 7nm process would be key to achieving these objectives.
Everything you need to know about RDNA 2 architecture in five keys
We already have a first look at the most important bases of the RDNA 2 architecture, but there are still some important questions to solve, and we are going to do it through a simple summary with five points that will allow you to know everything you need to know about this new generation of AMD.
Keep in mind, before reading on, that all the information that we are going to share with you below It has been confirmed in part by AMD. Those data that do not have that confirmation come from very reliable sources, and for this reason, it is most likely that they will be fulfilled. Now yes, we start.
1.-RDNA 2 will arrive at the end of the year
We do not have a defined date, but everything seems to indicate that they will be available between September and October. AMD will most likely roll out, which means that we should see the most powerful models first, and that in the coming months the mid-range and low-end models will be released.
Both NVIDIA and AMD have been following this approach for years, so we shouldn’t be surprised. If everything goes according to plan, Big Navi, AMD’s next top of the range based on the RDNA 2 architecture, should be the great protagonist of the presentation event scheduled for September or October.
2.-It will be an architecture to the last
And sincerely. In this sense, it is important to remember that the RDNA architecture, although it has shaped graphics cards with an excellent price-performance ratio, it is not prepared to work with the most advanced technologies in the sector, among which are functions as important as the set of optimization technologies that Microsoft bundled in DirectX 12 Ultimate.
This reality places the RDNA architecture in a clear disadvantage, in fact I have doubts about how it will age in the coming yearsEspecially if developers really start to take advantage of meshed shaders, variable rate shader, and sample feedback. RDNA 2 will be compatible with all of them, with the advantage that this implies.
3.-Hardware accelerated ray tracing
It is another of the most important innovations that RDNA 2 will bring to the architectural level. Graphics cards based on this architecture will have hardware dedicated to ray tracing acceleration, a feature that is already present since the end of 2018 in the RTX series 20 from NVIDIA.
I know what you are thinking, that AMD is late, but the truth is that everything depends on the prism with which you look at it. Ray tracing is going to take off with the arrival of next-gen consoles, and these won’t be available until the end of the year, so I personally think that actually has come at the right time. Says one owner of a RTX 2080 Super.
4.-Efficiency and performance go hand in hand
One of the big outstanding accounts that AMD had, even despite the improvements it introduced with the RDNA architecture, was boost efficiency of your graphics cards, that is, improve performance per watt.
According to the latest official information, the company led by Lisa Su has successfully achieved this objective, since the RDNA 2 architecture has, in theory, 50% higher performance per watt than the RDNA architecture. On paper it sounds good, but we’ll see what impact it has in the real world.
5.-AMD wants cheap prices
The Sunnyvale company has said it in many different ways, but the conclusion is the same. With RDNA 2 architecture they want to democratize 4K gaming and ray tracing, which means that your new graphics cards will be “affordable”.
So far so good, but what is an affordable price? We have nothing concrete on which to base ourselves, but I think it makes sense to think that AMD was doing a comparison with the prices of the 20-series RTX. If this is confirmed the story is told alone, the new generation Radeon RX will priced lower than RTX 20 series.
Endnotes: not a good time to buy a graphics card
We are awaiting the launch of two new graphics generations, NVIDIA’s Ampere and AMD’s RDNA 2. Both will arrive between September and October, so the best thing we can do right now if we need to buy a graphics card is to wait. In case you can’t do it, you have the option of going to the second-hand market and saving the day by buying something with a minimal investment that allows you to last a couple of months.
We do not have official information on the real performance that both architectures will offer, so it is impossible to anticipate which of the two generations will be more powerful. Multiple Sources Assume NVIDIA Will Beat Ray-Tracing Performance, and that thanks to DLSS 3.0 technology could leave AMD in a complicated position, but they have no greater validity than that of a simple rumor.
The same applies to AMD and the RDNA 2 architecture. Some say the company has managed to shape revolutionary architecture, and that its performance will not only be very good in ray tracing, but it will also be at the level of the NVIDIA RTX 3080 Ti. I see it complicated, but hopefully it is, more than anything because it would be something that would open a price war, and that would benefit consumers.