Big Navi means AMD is finally competitive •

Welcome to the red team. This is the first message to say hello when you lift the lid with one of AMD’s new Radeon 6000 series graphics cards, and it’s a good fit-after all, AMD fans are one of the most enthusiastic people in the game. It’s a department and has been expected by the entire “Big Navi” GPU community with almost savior enthusiasm. Thankfully, all of that positive and memorable energy has found a valuable target on two of AMD’s Radeon graphics team’s most powerful works, the RX6800 and RX6800XT.

Don’t worry before explaining how well these cards work in a wide range of games. plenty That-let’s easily set some expectations. For $ 580 (£ 530) and $ 650 (£ 600), respectively, the RX6800 and RX6800 XT slots are above the $ 500 (£ 450) RTX 3070 and below the $ 700 (£ 650) RTX 3080. Outperforming the RX6800’s RTX3070, the RX6800 is well suited for drawing on high-end Nvidia cards. More importantly, AMD is finally competing again at the high end of the graphics market. This is something the company hasn’t done since the unlucky Radeon 7 almost two years ago. Normally For many years than that.

These Big Navi cards not only offer significant performance improvements, but also mark the PC debut of AMD’s RDNA2 architecture, which forms the basis of both PlayStation 5 and Xbox Series X | S consoles. Like these next-generation machines, the RX 6000 Series supports hardware-accelerated ray tracing, ultimately ending Nvidia’s RT monopoly. In addition, it supports a number of other DirectX 12 Ultimate features that can improve performance in a variety of ways, including DirectStorage, variable rate shading, and mesh shaders. And when it comes to improving performance, AMD engineers have added features such as “smart memory access,” “rage mode” overclocking, and “infinity cache.” All of these should collude to push these cards to the limit.

From this embarrassing explanation alone, we can see that there is much to cover here. So let’s take a closer look at each of these cards. How are cards built, how are hardware stacked, and how are they tied to new features? Do you offer here?

This content is hosted on an external platform and will only be displayed if you accept targeting cookies. Please enable cookies to view them.

AMD GPU specifications CU Boost clock VRAM Mem interface TDP price
RX 6900 XT 80 2250MHz 16GB GDDR6 256-bit + 128MB IC 300W $ 999
RX 6800 XT 72 2250MHz 16GB GDDR6 256-bit + 128MB IC 300W $ 649
RX 6800 60 2105MHz 16GB GDDR6 256-bit + 128MB IC 250W $ 579
RX 5700 XT 40 1905MHz 8GB GDDR6 256 bits 225W $ 399
RX 5700 36 1725MHz 8GB GDDR6 256 bits 180W $ 349
RX 5600 XT 36 1560MHz 6GB GDDR6 192 bits 150W $ 279
RX 5500 XT twenty two 1845MHz 8GB GDDR6 128 bits 130W $ 199

The spec sheet can be a little boring, but if you look at it, you’ll see quite a lot this time around. AMD runs on the same 7nm TSMC process as the original “Little” navigation design, but the new architecture can deliver far more performance. This is driven by significantly faster clock speeds across the board and greater complementation of more individually performing computing units. There is also dedicated ray tracing hardware backed up for each CU. It is 10 times faster to calculate ray intersections than a pure software approach. It’s also compelling to compare with Nvidia’s 2nd generation RT hardware.

These computing improvements are underpinned by innovative new memory solutions. By pairing 16GB of GDDR6 video memory with 128MB of cache, the design is similar to the L3 cache on AMD’s Ryzen processor, and AMD has some fierce results from the card’s relatively pedestrian 256-bit bus. Can be achieved-slide, providing 2.17 times the bandwidth expected of a system standard 256-bit interface. This increase in memory bandwidth helps to offset higher compute unit (CU) counts and higher clock frequencies while controlling power consumption. This is a clever move, meaning that the early adoption of (expensive) GDDR6X memory in the RTX 3080 has at least partially improved Nvidia’s benefits. AMD also means that both models can offer 16GB of VRAM. This will surely save the nerves of gamers who have a large number of next generation games in their pipeline and are worried about running out of video memory. This 16GB of VRAM is also easily accessible from Ryzen 5000 systems via AMD’s Smart Access Memory tech. This is a performance enhancement feature detailed on page 6 of this review.

This completes the internal architectural improvements. What about external industrial design?

Custom RX6800 and RX6800 XT cards are manufactured by a wide range of AMD partners, but today’s review focuses on AMD reference models. These GPUs are set to be retailed on each card’s base RRP ($ 580 for the RX 6800, $ 650 for the RX 6800 XT), but with an impressive 3-fan axial cooler and plenty of metal. It feels like it’s not an entry-level design with heat-dissipating fins and an aluminum shroud that’s more complex than the box-shaped Radeon 7. Overall, AMD’s Radeon 6000 cards are heavy and very well constructed, with no give or flex found anywhere. The Vanilla RX 6800 is particularly dense with a more compact 2-slot design, and the RX 6800 XT’s 2.5-slot design provides more space for high-power cards to breathe. Both designs are the same length and height, 267mm (10.5 “) long and 120mm (4.7”) high, so they should fit comfortably in a standard ATX case.

These designs are a big step forward from the blower-style RX 5700 series cards and should perform much better in standard PC cases with proper airflow. I especially enjoyed the zero RPM fan mode. It seems to be operating from about 52 degrees Celsius and cuts to completely silence the fan during normal daily computing tasks. Of course, this is far from the first GPU to support such a feature, but it’s rare to see it in first-party designs, demonstrating the quality of AMD’s cooling solutions. Throughout my testing, the card worked quietly, and with the “balanced” power profile I used for the test, I limited the fan speed to just 50% of the maximum 2000 RPM. It’s up to Buildzoids and Gamers Nexus around the world to dig deeper inside, but the 15 power phases and 14-layer PCBs are certainly impressive-and we’ll take a closer look at power usage later on this page itself.

Regarding the port, we are considering clothing that is slightly different from Nvidia’s RTX 30 series card. In addition to one HDMI 2.1 port (ideal for next-generation 4K 120Hz or 8K 60Hz TVs you might have picked up for use with new consoles), two DisplayPort 1.4 ports and one USB-C port (of course) Can be used as a third DisplayPort, HDMI, or DVI-D with the appropriate adapter). Like Nvidia cards, it includes AV1 decoding, allowing sites such as Netflix and Twitch to stream high-quality, high-frame-rate video content while requiring significantly lower bandwidth than current codecs. I will. Finally, this card is compatible with PCIe 4.0, but it also works fine with PCIe 3.0 slots. For power, both cards have two 8-pin inputs. AMD recommends a 650W PSU for the 6800 and a 750W PSU for the 6800XT. This is very suitable for midrange to high end graphics hardware.

We used Nvidia’s (agnostic) PCAT system to more accurately determine the power efficiency of Big Navi during the game. This is an interposer board that is placed between the graphics card and its power input to measure actual power consumption. Combined with in-game performance metrics such as frame rate and frame time, you can accurately calculate the wattage required to render each frame, so you can measure actual efficiency. AMD states a 54% improvement in performance per watt between RDNA and RDNA2, is that supported by testing? And how will the current efficiency champion Nvidia’s Ampere architecture be compared?

Jules per frame RX 6800 XT RX 6800 RTX 3080 FE RTX 3070 FE RTX 2070 RX 5700XT
Death Stranding 2.933 2.644 3.349 2.915 3.651 3.846
Percentage difference 110.9% 100% 126.6% 110.2% 138.1% 145.5%
Jules per frame RX 6800 XT RX 6800 RTX 3080 FE RTX 3070 FE RTX 2070 RX 5700XT
Gears 5 4.384 3.792 4.156 3.603 4.734 5.403
Percentage difference 121.7% 105.2% 115.3% 100% 131.4% 150.0%

In Death Stranding, AMD’s RDNA2 architecture is slightly more efficient than Nvidia’s Ampere, significantly superior to first-generation RDNA (represented by the RX 5700 XT), and AMD claims increased efficiency. Is approaching. However, this game is very AMD hardware friendly, so what about more balanced options?

In Gears 5, the 3070 has proven to be a competitive and slightly more efficient card with the RX 6800, but with a margin of only 5% in favor of Nvidia. Obviously, both new architectures are significantly more power efficient than the previous ones, demonstrating the engineering built into each option.

Interestingly, the RX 6800 XT requires 15-20 percent more joules per frame in the two games tested, so the more compact 6800 is arguably the series leader in terms of power efficiency. This is almost not unexpected, as we saw similar margins between 3070 and 3080.

After a quick power test, it’s time to benchmark the game. If you haven’t checked the RTX 30 Series reviews yet, you’ll see some new titles and new test platforms here. If you read those reviews-thank you! -There is nothing new here, so go to the next page.

We tested AMD’s Ryzen 3950X as a candidate for a GPU test bench build. This gave us access to all the PCIe 4.0 bandwidth supported by AMD’s 6000 series and Nvidia’s 30 series, but eventually we chose the Intel Core i9 10900K for our test system. , Replaces the Core i7 8700K used in previous generation tests. Intel systems are technically limited to PCIe 3.0 bandwidth, which can limit performance in GPU-bound scenarios, but on Intel chips than the fastest Ryzen processors available at the time. I got a high 4K frame rate. The Intel system showed more important frame rate reads at 1080p and 1440p, so it’s the best way to avoid CPU bottlenecks as much as possible and allow different graphics cards to be distinguished. I’ll consider the Ryzen 5000 or its successor the next time I update the testbench, but for now it makes sense to stick with it, so re-dozens of graphics cards during the busiest time of the year. No need to test (during a new console cycle).

To allow the results to be compared, the Core i9 10900K is locked to all core frequencies of 5GHz and cooled by the 240mm Alpacool Eisbaer Aurora AiO to keep the overclocked system at about 75C at full load. The 10900K is backed by an Asus Maximus 12 Extreme Z490 motherboard and two 8GB sticks, the G.Skill Trident Z Royal 3600MHz CL16. Our game runs from the large capacity 2TB Samsung 970 Evo Plus NVMe drive provided by Box. The entire rig is powered by a 850W gold rated gamerstorm power supply.

Now that you’re familiar with the testbed and have an overview of the card itself, let’s start some game tests to see the performance of the card.

AMD Radeon RX6800 and RX6800XT Analysis

  • First, hardware and power analysis [This Page]

  • Doom Eternal, Control, Borderlands 3, Shadow of the Tomb Raider-Game Benchmark Part 1

  • Death Stranding, Far Cry 5, Hitman 2, Assassin’s Creed Odyssey-Game Benchmark Part 2

  • Metro Exodus, Dirt Rally 2, Assassin’s Creed Unity-Game Benchmark Part 3

  • Control, Metro Exodus, Battlefield 5-RTX Game Benchmark

  • Smart Access Memory Benchmarks and Requirements

  • RX6800 and RX6800XT-Digital Foundry Verdict

Back to top button