The long-awaited AMD Big Navi, RDNA2, GPUs have finally been announced, and they are looking to be a powerhouse of gaming performance. Back when Nvidia had first announced their GeForce RTX Ampere 3000 Series GPUs, many believed that their prices were so competitive because AMD had their own big GPU announcements to make – and they were right.
Not only are these three new AMD RX Radeon 6000 Series GPUs on the same level as their chief competitor’s counterparts, but, in some cases, they actually surpass their performance, despite being more energy efficient, and having a lower price. The only concern for these GPUs is how sustainable their benchmarks are, as AMD has faced issues with high GPU temperatures in the past. One such example is the Radeon 5700 XT, which had incredible (for its price) benchmarks, but its overheating issues ultimately devalued its performance. This time around, AMD has announced that these Radeon 6000 series GPUs will have twice the performance per Watt of their predecessors, and their cooling solutions appear to be far more sufficient – though only benchmarks will truly show how sustainable their performance will be.
Additionally, AMD announced some excellent new features, which are targeted towards rivaling Nvidia’s own equivalences. For example, AMD has implemented the second-generation latency reducing technology named AMD Radeon Anti-Lag which, together with Radeon Boost, can reduce end-to-end input lag, similar to Nvidia’s Reflex – though in some cases (like 360 Hz monitors) not as effectively. Additionally, through the synergy between AMD’s CPUs and GPUs, AMD has announced their Smart Access Memory, which, together with the “Rage Mode” – a one click overclocking feature – can provide a boost in performance ranging from 2% to 13%, depending on the title, at a 4K resolution.
The 7 nm RDNA processing node has allowed for both higher performance Compute Units (CU) and has provided the groundwork for AMD’s Infinity Cache. These new CUs have redesigned data paths, a new pipeline rebalancing, and pervasive fine-grain clock gating, making them 30% more energy efficient. The Infinity Cache, which is based off AMD’s Zen 3 Level 3 Cache, allows their 256-bit memory bus to delivery more than twice the bandwidth of a 384-bit bus, at a lower energy expenditure. With this new generation of GPUs, AMD seems to have matched Nvidia’s RTX 3000 Series graphics cards’ performance, while having both a smaller TDP, as well as a smaller form factor.
Three were the announced RDNA 2.0 GPUs: The Radeon RX 6800, the Radeon RX 6800 XT, and the Radeon RX 6900 XT, all of which are meant to directly compete against the equivalently priced Nvidia GeForce Ampere RTX 3070, RTX 3080, and RTX 3090. In addition, AMD went on to provide a list in which they pair each of their new Radeon GPUs with one of their Zen 3 Ryzen 5000 Series CPUs during their presentation. This infographic pairs the Radeon RX 6800 with the Ryzen 7 5800X, the Radeon RX 6800 XT with the Ryzen 9 5900X, and the Radeon RX 6900 XT with the Ryzen 9 5950X. That being said, seeing as how it appears that the Ryzen 5 5600X, and the Ryzen 7 5800X are poised to become the best, and most cost efficient, CPUs (specifically for gaming applications), this list might not be the most efficient. Any of these three new GPUs can be paired with either Ryzen 5 or Ryzen 7 processors but, for the majority of gamers, the Radeon RX 6800 and Radeon RX 6800 XT will be the most suitable options for a new gaming rig.
For any such build, the GPU will most likely turn out to be the most expensive component of the setup, so the need for it to provide optimal performance for its intended use is paramount. Choosing between these two new Navi 21 GPUs is tricky, as they both seem to provide incredible performance for their price; especially in terms of how they rival their Nvidia GPU counterparts. Therefore, choosing between the two not only depends on how they fair against each other, but also on the value that they provide versus their competition, as Nvidia’s GPUs are always a valid option. Additionally, how they fair against their competition’s price points will ultimately decide which GPU has the better price-to-performance ratio.
To ascertain which GPU is best for your intended use and budget, we will compare these two mid/upper-midrange GPUs, both in specifications and benchmarks, and distinguish whether a clear winner arises, and what the optimal use for each GPU will be.
RX 6800 vs RX 6800 XT: Specifications
Radeon RX 6800
Being the new chief competitor to the GeForce RTX Ampere 3070, the Radeon RX 6800 is equipped with a Navi 21 XL GPU, using a TSMC 7 nm processing node that has 26.8 million transistors on a 536 mm2 die. It has 16 GB of GDDR6 VRAM, with a 256-bit memory bus, a 128 MB Infinity Cache, and a bandwidth of 512 GB/s. In terms of rendering power, it features 3,840 shader units, 240 TMUs, 96 ROPs, 60 ray tracing cores, and 60 CUs; providing a FP16, FP32, and FP64 performance of 32.33 TFLOPs, 16.17 TFLOPs, and 1.01 TFLOPs respectively. The base clock speed of this GPU lies at 1,372 MHz, which can be partially boosted to 1,815 MHz (Game Clock speed), and fully boosted to up to 2,105 MHz. The Memory Clock speed is 2,000 MHz and 16 GB/s. Finally, the board will have a two-slot width, and a TDP of 250 Watts. The Radeon RX 6800 will have a baseline price of $579, and will launch on November the 18th, 2020.
Radeon RX 6800 XT
Next, we have the equivalent to Nvidia’s GeForce RTX Ampere 3080: The Radeon RX 6800 XT. This mode will use the Navi 21 XT variant, which has the same foundry, processing node, transistor count, and die size of the aforementioned Navi 21 XL variant. Additionally, the VRAM quantity and quality will also be identical to that of the Radeon RX 6800, but when it comes to the rendering configuration, this GPU has 4,608 shading units, 288 TMUs, 128 ROPs, 72 RT Cores, and 72 CUs, delivering an FP16, FP32, and FP64 power of 41.47 TFLOPs, 20.74 TFLOPs, and 1.296 TFLOPs respectively. The clock speeds of this reference card are quite impressive, as it has a base clock speed of 1,487 MHz, a Game Clock speed of 2,015 MHz, and a Boost Clock speed of 2,250 MHz. Like the RX 6800, it has a Memory Clock of 2,000 MHz and 16 GB/s. This GPU will have a 2.5-slot width, with a slightly increased TDP of 300 W. It will sell for $649, and will be released together with the aforementioned RX 6800.
RX 6800 vs RX 6800 XT: Comparison
So, what does the Radeon RX 6800 XT offer that justifies its $70 price increase? Firstly, it has 768 more shader units, 48 more TMUs, 32 more ROPs, 12 more RT cores, and 12 more CUs; which allow for an increase in FP16, FP32, and FP64 computing power of 9.14 FLOPS, 4.57 TFLOPS, and 0.286 TFLOPS respectively. Additionally, the GPU will have better overall clock speeds, with a higher base frequency by 115 MHz, a higher Gaming Clock frequency by 200 MHz, and a higher Boost Clock speed by 145 MHz. The downsides to these inherent advantages are that it will have a larger form factor, and will require 50 additional Watts of power in order to operate. If form factor is not an issue for your build, then its specification advantages are well worth its additional price.
AMD has released benchmarks comparing two GPUs to two different models: They’ve compared the RX 6800 XT to the GeForce RTX Ampere 3080, and the RX 6800 to the GeForce RTX Ampere 2080 Ti (most likely due to the fact that the RTX Ampere 3070 has not yet officially been released). Do note, that there is a chance that many of these benchmark values may be cherry picked by AMD, and therefore they are to be taken with a hint of skepticism.
Radeon RX 6800 vs RTX 2080 Ti
Beginning with the latter comparison, we see that the Radeon RX 6800 (with Smart Access Memory enabled – which requires a Ryzen 5000 Series CPU) we see that the RX 6800 leads the RTX 2080 Ti in 4K gaming by anywhere between 2 and 30 FPS, depending on the title. This advantage grows significantly at a lower, 1440p, resolution, where we are told it outperforms the RTX 2080 Ti by upwards of 50 FPS in Forza Horizon 4, and has an overall 18% performance lead on average.
In regards to how well the Nvidia RTX 3070 compares to the RTX 2080 Ti, we have an article which shows that the RTX 3070 will most likely have a slight edge on the RTX 2080 Ti at 1080p or 1440p resolutions, which will all but disappear at 4K.
All in all, the Radeon RX 6800 is looking to be a more powerful alternative to the RTX 3070 – which is also why it costs $79 more at launch – and have double its amount of VRAM.
Radeon RX 6800 XT vs RTX Ampere 3080
When comparing the Nvidia GeForce RTX 3080 with the Radeon RX 6800 XT, we see that the AMD GPU has an advantage in five of the ten different titles that were tested. Of course, all these framerate discrepancies are rather small, so overall it is fair to say that the two GPUs perform on the same level when it comes to 4K resolution gaming. Where the Radeon RX 6800 seems to outperform the RTX 3080 (though by slightly) is at 1440p, where it holds an advantage on eight titles of the ten titles. Additionally, with the use of AMD’s Smart Access Memory, these advantages will increase by anywhere between 2% and 13%, so it is almost a no-brainer for those looking to purchase a Ryzen 5000 Series CPU for their build.
Considering the fact that the RX 6800 XT has 6 GB more VRAM than the RTX 3080 (though GDDR6, unlike the GDDR6X of the RTX 3080), and the fact that it costs $50 less while still being more energy efficient, make it an overall better choice to pair with an AMD Ryzen 5000 Series CPU. Indeed, if AMD’s benchmarks turn out to be accurate and no other distinct disadvantages arise for the Radeon RX 6800 (like overheating), then AMD will have truly created an overall better GPU than that of Nvidia.
Depending on what you’re looking for from your future gaming build, the answer in regards to which GPU is better for you will vary. If you are looking to play competitive online multiplayers at the highest level, then perhaps neither of these two GPUs will be an optimal choice. Nvidia has teamed up with five different monitor manufacturers, in order to provide a series of 360 Hz monitors that, when paired with Nvidia Reflex, can significantly decrease end-to-end input lag by a substantial amount, as well as display overall input lag in real time. Of course, these monitors do cost anywhere between $700-$1200, so they are mostly recommended solely for gaming professionals. If you aren’t that serious about competitive gaming, but would still like to play at high 1080p framerates, then the RX 6800 will surely suffice.
For those that enjoy 1440p 144 Hz gaming, and want consistent 144 Hz framerates with maximum in-game quality settings, then the RX 6800 XT is the way to go. Its extra cost is well worth the money, considering how next-generation games will become more and more hardware demanding as time goes by. Of course, the fact that the Xbox Series X and Playstation 5 also have RDNA2 based GPUs, means that game developers will most likely optimize their titles for this architecture.
If you love single player triple-A titles, and would prefer to play them at a 4K resolution at 60 FPS, then the RX 6800 appears to provide the best value for its cost. It hits, and exceeds, the 60 FPS score for every game AMD tested, and will most likely continue to do so for next-generation games, as most titles will be optimized for consoles, which have claimed to be able to achieve 60 FPS at a 4K resolution, despite being less powerful than the RX 6800.If you are planning on building a new PC, and are excited to use one of AMD’s new Ryzen 5000 Series, Zen 3, CPUs that will launch on November the 5th, then AMD’s new GPUs are an excellent option for your new setup… Given that AMD doesn’t fumble their launch similar to Nvidia’s RTX 3080 and RTX 3090 release. Perhaps it is time for us to witness a transition toward the reign of a new GPU king, as this series of GPUs definitely pus AMD a step closer toward seizing Nvidia crown.