Skip to content

6900 xt vs rtx 3090

November 17 Update: Buy the AMD Radeon RX 6900XT here.

November 13 Update: One day closer to us, the search continues to find the list of Radeon RX 6000 series GPUs.

November 12 Update: After evaluating the Ryzen 5000 CPU, we are eager to get down to business I got in touch with the RX 6000 GPU to understand the full potential behind intelligent access memory.

Nov 11 Update - Our team is constantly looking for links online, so be sure to check this page regularly as they may appear anytime we find them. We're just days away from launch, so the excitement is starting to heat up.

Following AMD's big reveal on 6000 series GPUs on October 28, one of the most interesting topics is how the 6900 XT (commonly referred to as "Big Navi"), the most powerful graphics card in the series, will compete with Nvidia. product, RTX 3090. For now, we can only start with the first leak of official data and independent benchmark tests provided by AMD at the time of its release, but we will try to see how the red team’s challenger compares to the green team’s heavyweight champion (in most of them) Without permission). .

If you are interested in seeing a more general comparison of the two ranges as a whole, please check this article.

As we will discuss, not all sub-components of the two cards can be directly compared (including GPU core, ray tracing core, Tensor core, and memory category), but the following table provides us with an overview of how they are measured. on the paper.

For those running higher resolution displays, especially 4K, the memory bottleneck proved to be an increasingly serious performance issue. Regardless of the GPU clock speed and core power, if the memory bandwidth is not enough, this will usually result in a lower minimum FPS for comparison, even if the average FPS is high, which leads to potential lag in the game because the frame rate is significant. fluctuation. The

6900XT is inferior to the Green Team’s flagship model RTX 3090 in terms of VRAM, video memory speed, and video memory bus; however, AMD has developed a secret weapon. Based on the technology developed for its Ryzen CPU, AMD’s "unlimited cache" clearly allows the GPU to run at 2.17 times the bandwidth, which is usually the capability of a 256-bit memory bus. This means that according to AMD, its card can run at an effective memory bandwidth of 1,664GBps, which is almost 1.8 times the memory bandwidth of the RTX 3090. This means that 6900XT should outperform Nvidia cards in certain areas where a smaller bus is used. Save energy consumption and heat generation, and help reduce the TDP and TBW of AMD cards.

Historically, software and drivers have been an area where AMD lags behind Nvidia, although the 6000 series seems to be closing this gap.

AMD and Nvidia cards are fully compatible with "DirectX 12 Ultimate": this means they will support mesh shading, variable speed shading, and sampler feedback. DLSS (Deep Learning Super Sampling) is an advantage of RTX 3090 owners over their competitors. This is an exclusive proprietary technology that Nvidia is unlikely to provide to its competitors in the short term. Of course, this is only a factor. . Those games that use this technology. Nevertheless, for this reason alone, RTX 3090 currently has the upper hand in 8K games, even though AMD has revealed that they are developing their own version of artificial intelligence. The enhancement technology called Super Resolution, we don't know when it will be available to 6900 XT users, and we don't know how effective it is.

Raytracing (aka RTX, where the Nvidia card prefix comes from) is another area where Nvidia has the upper hand, although AMD has announced that the 6900 XT will use its own version of technology when it is released. However, according to Videocardz, early testing showed that AMD's implementation still lags behind Nvidia's.

In addition to the aforementioned Rage mode and smart memory access, 6900 XT users will also use the new "Radeon Boost" and "Radeon AntiLag" technologies to increase response time and reduce latency. In their CPU and GPU versions, AMD has indeed promoted the latency reduction aspect of their technology, and it may be an area where they think they can truly surpass Nvidia. We hope to see third-party data in the near future.

AMD did not give us the traditional "base clock" speed measurement, but chose to use their "game clock" measurement, which should be an indication of the speed that the GPU expects to achieve during a typical game, not an extremely low speed. Even an idle computer rarely reaches the level represented by the "base clock". At this stage, we cannot directly compare the two cards on the base clock or the game clock, because Nvidia does not use the latter method. In any case, leaving the Boost Clock number, we can only assume that the AMD card is ahead of all these indicators.

However, clock speed does not necessarily translate into performance: factors such as the number and power of cores and the resulting IPC (instructions per clock), as well as other indicators, such as the number and efficiency of ROP (units that process output). The rasterization process (converting basic vector information from the CPU into a fully developed texture) is combined to produce real-world performance.

Unfortunately, we cannot directly compare the two metrics used by the two companies for their GPU cores (the AMD core stream processor and Nvidia’s CUDA), because each metric is calculated differently. Similarly, "Ray Tracing Core" works differently on the two cards, and "Tensor Core" is a name that only Nvidia actually uses.

The benchmark test we must use is provided by AMD as part of its important presentation, which only covers the performance of these two graphics cards at 4K resolution. However, this is the design screen of these two cards, and it is also the theater where Nvidia is expected to dominate. If you are not interested in operating a 4K monitor (or multiple 1440p monitors), we recommend that you consider a 6800 XT or RTX 3080 (if you can).

What is striking is that AMD only shows FPS performance in a variety of games with Rage mode enabled and smart memory access. We know that in Forza Horizon 4, their performance is the best than RTX 3090. These two activated functions contribute an additional 13% FPS increase, which means that if there is no one, we can expect this 150 FPS or more is seen on the game. game. On the bar graph, this is still higher than the RTX 3090 (assuming it scales proportionally), but this is AMD's best performing game. For other games, while the FPS boost will almost certainly vary from game to game, we can expect RTX 3090 to lead.

It turns out that the most important factor in determining whether people choose AMD or Nvidia is how easy it is for them to buy either card. In theory, the previously released 3090 RTX should give you a head start in attracting potential customers to this price card, especially since the target market for those who are willing to pay for the enthusiast card premium is much smaller than that of the card. two ranges The value of the low input. However, Nvidia's 30 series launch has been plagued with inventory issues. As is well documented, according to Nvidia's announcement on the matter, this situation does not appear to change until spring 2021. This may be a breakthrough that AMD needs to change. The staff converts from the green team to the red team.

However, there are many questions about whether people who buy the AMD 6000 series will face the same problems. It may be that AMD has accumulated a higher inventory when it was first released, but given that we see unprecedented demand for GPUs this year, our money is that the inventory will soon be exhausted.

In the end, your decision to use AMD or Nvidia flagship card depends on your situation. If you want to assemble a new version and consider buying a new Ryzen 5000 series, the performance improvement that smart memory access and Rage mode will bring to you should be very attractive-providing 3090 performance comparable to Nvidia RTX for less than $500. If you already have an older Intel or AMD CPU and do not plan to upgrade, then the decision is less clear. NVIDIA’s RTX 3090 is undoubtedly the best performance, especially if you want to play games that support ray tracing and DLSS, if it’s worth the extra $500 to give you the best performance, then NVIDIA’s flagship seems to be suitable Your one. There are two caveats to all of the above: First, if you can wait for the independent benchmark test of the new AMD GPU (after December 8 of the 6090 XT), and then make a choice, we strongly recommend that you do so. Production; secondly, whether you can really buy any card from the dealer in time according to your needs, this may eventually become the deciding factor between the two.

Previous article Flight simulator intro games hardware

Leave a comment

Comments must be approved before appearing

* Required fields