Nvidia RTX vs. AMD Big Navi — The Battle For Graphics Performance
The two biggest players in the GPU industry are Nvidia and AMD. It seems that Nvidia is preparing a new RTX card to compete against AMD’s Big Navi, after reports that it is going to be more powerful than competing products. This is a big deal since the market is a niche that doesn’t have many producers. Whoever can release the better product can capture their lion’s share and more. According to reputable leaks, Nvidia is going to answer that with a boost to the RTX 3070 Ti. The flagship the Big Navi (AMD RX 6800 XT) is reportedly beating in performance would be the RTX 3090, but it is not absolute and based more on leaked reports.
Nvidia RTX Series
The RTX series brings Nvidia’s vision of ray tracing to light. This is a rendering platform that brings virtual realism in graphics at a higher cost, but with the best results. It is a premium product that is aimed for serious imaging products ranging from games to simulations. It brings photorealism by using techniques that simulate how light would interact with virtual objects. This is great for creating virtual worlds that mimic reality for gaming and entertainment applications.
The RTX is built from Nvidia’s Ampere GPU microarchitecture and Turing architecture which utilize Tensor and RT cores for ray tracing acceleration. The RTX drivers are available through Optix and DirectX for users. RTX also includes AI integration (Deep Learning Super Sampling DLSS), common asset formats, rasterization (CUDA) support and simulation APIs. The GeForce RTX is Nvidia’s product line, with the high end 30 series being its flagship.
In terms of performance, Nvidia’s DLSS feature boost frame rate using dedicated AI processing Tensor Cores on the GeForce RTX. This also offers lower latency and faster response, so that the rendering is smoother and not jerky, lagging or chopping up. Users can utilize this for the maximum fps (framers per second) to produce the highest quality images.
AMD Big Navi
AMD claims that their Big Navi will be:
‘by far the most powerful gaming GPU we have ever built’.
That raises the bar much higher for Nvidia. The Big Navi is a part of the Radeon graphics card product line marketed as the RX 6000 series. One of its main feature is 4K UHD gaming at over 60 fps. At the moment there has been very little detail of the full technical specifications released about Big Navi, other than reports from AMD and other reviewers. The stories about its alleged performance has been from reputable news leaks. Come October 28, 2020 more news about Big Navi will be officially released by AMD.
I can assume that the Big Navi is based on the RDNA core graphics architecture, which is used in the current Radeon 5000 series. For the RX 6000 series, it will probably be using RDNA 2. For this to be an RTX killer, it would surely have to surpass the performance indicators. Among the applications that can run on AMD’s RDNA 2 include Unreal Engine 5. The RDNA 2 graphics chips will also be used by Sony PS5 and Microsoft’s Xbox Series X, so this has made an impression with game console developers.
According to Tom’s Guide, there is serious indication that the Big Navi will be much better than Nvidia’s RTX.
YouTube channel Moore’s Law is Dead has uncovered information that claims a Big Navi GPU will deliver 40 to 50 percent more performance over a GeForce RTX 2080 ti, Nvidia’s current top-end graphics card.
… two clusters of compute units measuring in at 36CUs each. A total of 72CUs translates to 4,608 stream processors, which when combined with a 384-bit bus, GDDR6 video memory, and a maximum clock speed of 2,510MHz …
The best way to make a comparison at the moment is from the released benchmarks that have resulted from initial testing.
The Benchmarks
Performance on paper tends to differ from actual gaming and rendering, from the user’s perspective. Benchmarks determine this by testing the product with applications. The benchmarks are all initial tests, but do show results to expect from the final product after it will be officially released.
Another source for a benchmark test can be found at igor’sLAB.
What is interesting in the benchmark results is the 4K test using Firestrike Extreme show consistency in Big Navi beating out the RTX. However, when it comes to ray tracing, the RTX may have the advantage. The scoring also differed between the two benchmark sources, with igor’sLAB metrics using percentage while the Yuko Yoshida testing converted values to benchmark scores. Another difference is that the Yuko benchmark used more RTX series products, while igor’sLAB only tested two RTX (not the RTX 3090).
I won’t say that these benchmarks are a 100% reliable or accurate way of measuring the performance. Until we have the final release of the products, we can come up with more concrete observations. At the moment take these benchmark results as just an idea of the performance. We also have to factor in the type of CPU that was used and the testing environment. That will at least give us more idea of the type of machine that was used and how we can replicate it for our own benchmark testing.
Looking Towards The Future
While the test shows how much higher the Big Navi can reach in 3DMark scores, it doesn’t outperform the RTX in all the tests. Nonetheless, if the data is accurate and verified, it is still significant indicators. Perhaps we need to see more tests using optimized games for ray tracing and 4K to see overall how the Big Navi stacks up to the RTX line.
Another factor that could determine overall performance is the fabrication process for the circuit design. Big Navi will continue using the 7 nm process from the RDNA chip, which increases the transistor count to 21 billion (codename Navi 21). Nvidia also uses a 7 nm process for their Ampere-based GPU, but packed with 54 billion transistors. In terms of thermals, with smaller transistors they may run faster, but also generates more heat. It also consumes more power, as these chips process data. Thermals may be something to watch out for in case issues of overheating begin popping up.
The early benchmarks were not official test results, but based on early testing done independently. I cannot verify the data, but it shows the potential for Big Navi’s performance. There are even reports that Big Navi may not turn out to be the RTX killer. Overall this should enable users to make use of the offerings from advanced AI and ray tracing acceleration in the next generation of graphics intensive applications as the demand grows for compute power. Usually AMD underperforms when compared to Nvidia, but it looks like the gap is closing.
