Right now, we’re arguably in the most exciting battle for GPU superiority that I can remember. Nvidia has recently unveiled its RTX 30-series lineup, boasting performance benchmarks that destroy their 20-series lineup in both gaming and productivity workflows. On the other side of the GPU pond, AMD has just unveiled more intricate details regarding their RADEON RX 6000-series (BIG NAVI) GPU lineup, with initial benchmarks suggesting that team red could be firmly in the driver’s seat (when referencing gaming performance) for the first time in over 7 years. I know, crazy.
With everyone trying to digest this breaking news, many are now asking the big question, what’s better; AMD’s RADEON 6000 series or Nvidia’s 30-series?
In the following article, we’ll be taking a closer look at the specs, benchmarks, and overall value that both RADEON and Nvidia are bringing to the table. We’ll be doing side by side comparisons with both inhouse benchmarks (3080/3070) and pitting them against what AMD has shown during their recent keynote.
So, with a ton of data to get through, let’s waste no further time and dive straight into it!
The battle between Nvidia and AMD dates back a good while now, with AMD losing out in the sales market for as long as I can remember. As we look back over the last 10-15 years, Nvidia has pretty much dominated the GPU sales market, despite not always having the best GPUs at a particular time.
Take 2008 for example, Nvidia had just released there GeForce GTX 260 and 280 cards – two fairly competent gaming GPUs. At this time, AMD had released their much smaller, cheaper, and equally as powerful RADEON HD 4870. On paper, this bout was a no brainer. The
AMD card offered better performance in most titles and was less power-hungry and cheaper. So, you’d expect AMD to have the dominant share when it came to sales that year – but they didn’t.
Even with the advancements the HD 4870 offered over both the 260 and 280, Nvidia still took the lion’s share of the market. This theme continued for the next 6-8 years, with AMD eventually accepting the inevitable – Nvidia would always sell more GPUs no matter what.
Fast forward to the mid-2010s, and AMD had stopped competing with Nvidia on a GPU to GPU basis. They decided to seek guaranteed cash over guaranteed losses by attacking the APU and console marketplace. It wasn’t until much later, when AMD starting to churn some real profit that they tried to tackle the gaming GPU marketplace once again.
Over the past couple of years, AMD has been getting ever closer to Nvidia when it comes to both gaming and productivity workflows. With their latest offerings bridging the gap in the mid-range GPU market.
Fast forward to the present day, and AMD’s RX 6000-series finally seems to be ahead of Nvidia for the first time since around 2012. Speaking of which, let’s take a closer look at some of the individual comparisons to see what each next-gen AMD/Nvidia GPU has to offer.
On paper, it’s clear to see the differences between both brand’s leading flagship GPUs – each offering wins in a number of different categories.
Whilst we can’t really compare the two GPUs in fields such as core counts (thanks to the way each architecture works), there are some useful comparisons that can be drawn when looking at clock speeds and memory. In terms of raw memory capacity, the RTX 3090 brushes the RX 6900XT aside, offering up 24GB GDDR6X over AMD’s 16GB GDDR6. The Nvidia also offers greater memory speed and a wider memory bus than the RX 6900XT. However, like in true AMD fashion, it isn’t quite that simple.
Thanks to AMD’s newly implemented Infinity cache (128MB), the 6900XT effectively has a memory bandwidth of 1,664GBps – almost double that of the RTX 3090. Whilst it’s unlikely that the 6900XT offers the full potential of 1,664GBps (shown by AMD’s use of the word ‘effective’), it’s still possible that AMD’s card will offer greater memory bandwidth in a lot of scenarios.
Briefly looking at clock speeds, the RX 6900XT has an easy win here, coming to the table boasting a higher base, boost, and game clock than its alternative. So, how does this all stack up in real-world scenarios? Well, let’s take a quick look at some gaming benchmarks.
As you can see, the RX 6900XT is right up there with the RTX 3090 when it comes to gaming at 4K resolution. It’s a coin toss between the two brands for which takes the lead, with AMD winning in 50% of the games. The other 50% have a very closely matched battle between the two cards – a big improvement over the previous generation AMD GPUs.
So, how does this all stack up in the pricing department? Well, as you’d expect, the Nvidia comes equipped with a higher price tag – big surprise. However, what might come as a surprise is just how much more expensive the RX 3090 actually is.
The RX 3090 currently retails for $1,499, making it a whopping 50% more expensive than AMD’s RX 6900XT ($1,000). That’s a massive difference for what is considered a very small performance increase over a couple of titles.
Like the 6900XT vs 3090, it’s fairly difficult to draw side by side comparisons when looking at the 6800XT vs the 3080. That being said, there are some comparable features and specs that do offer some eye-opening insights.
Firstly, we have the memory capacity. The RX 6800XT comes to shelves offering the same 16GB GDDR6 VRAM as the 5900XT, giving it a clear advantage over the RX 3080 (10GB GDDR6X) in this department. Having said that, Nvidia has equipped its 30-series GPUs with the much newer, more efficient GDDR6X memory – offering a wider memory bus than the RX 6800XT. This, in turn, gives the RTX 3080 a greater standard of memory bandwidth when compared to the AMD.
Like the 6900XT, the 6800XT has a trick up its sleeve – the 128MB Infinity cache. Thanks to this, the RX 6800XT has an effective memory bandwidth that far exceeds that of the 3080 – by almost double.
Having a look at clock speeds, the trend continues with AMD outperforming the Nvidia alternative by some margin. In boost clock frequency, the AMD has over 500MHz additional speed to play with – something that certainly plays its part in gaming scenarios.
As you can see, when comparing the 6800XT side by side with the 3080 at 1440p, the AMD offering almost always comes out on top. The only losses to be found in this particular field are in Division 2 and Resident Evil 3 – a huge feat for AMD.
This being said, these are the AMD slides taken from the recent announcement, many of which have been discredited due to the usage of both RAGE mode and Smart Access Memory – both of which we’ll touch upon in more detail shortly.
With that in mind, how do these two powerful GPUs stack up in terms of pricing and availability? Well, once again, the AMD option is a slightly cheaper proposition, hitting shelves at $650. That’s a $50 reduction when compared to Nvidia’s RX 3080 card. Whilst it’s safe to say that Nvidia will probably still offer the slightly faster GPU at this particular price point, it certainly gives most consumers something to think about, especially when you factor in the recent failure to produce enough GPUs to cover demand.
Lastly, we find ourselves at the lower end of the price spectrum. However, whilst the lower end of the price spectrum might bore some individuals, it’s at this price point where we’ll probably find the most competition.
Once again, comparing these two cards side by side is less than ideal, with many of the specifications only being relevant to either card – thanks to the architecture that they come equipped with. However, like the other theoretical rungs in this guide, we can take a closer look at clock frequencies, memory, and price.
Starting with clock speeds, once again, the AMD card comes out on top – albeit not by much. The AMD card offers a might clock boost of 2,015MHz which, when compared to the max boost of the Nvidia card, is pretty impressive. Even at boost clock, the AMD card is still clocked at almost 100MHz higher than Nvidia’s alternative.
Like all cards in this guide, the AMD GPU holds a slight advantage on process node. That being said, Nvidia has been catching up in this department ever since Turing generation’s 12nm mode.
Taking a closer look at memory differences between these two cards, we see a landslide victory for the AMD GPU. It comes equipped with the very same 16GB GDDR6 that both the 6800XT and 6900XT offer, alongside an equal 128MB infinity cache as well. Furthermore, the AMD card boasts victories in memory bandwidth and speed, whilst having the same 256-bit bus.
So, what does that all mean from a gaming standpoint? Well, we’re unsure at this point – we are running our own exclusive benchmarks on the 3070 as we speak. But for now, we do have a good indication of the performance differences based on 2080Ti benchmarks.
The RX 6800 sees clear wins in almost every title that AMD showcased during their live announcement. Granted, these will be best case scenario results, but they still give us a great indication of how the RX 6800 performs against their bitter rivals.
Amongst all the excitement that was the RX 6000-series launch, a few new terms were unveiled that left more than just a few consumers scratching their heads – Smart access memory and Rage mode.
These two features played a major role in all the benchmarking slides during AMD’s keynote which left many people speculating what they were.
Let’s start with Rage mode.
This is a new form of overclocking, a simple button within the RADEON software suite which, when turned on, allows you to unlock the maximum potential of the card by utilizing the remaining overclocking headspace. For all intents and purposes, this is another overclocking step that was previously unheard of.
Next, we have Smart access memory, this is a little more technical than rage mode and, to my knowledge, should have a greater impact on overall performance.
Smart access memory is a new feature introduced by AMD which allows their latest CPU lineup to take full advantage of the RX 6000-series infinity cache – instead of 256MB segments that have previously been the norm. In the simplest terms, smart access memory is a more efficient way of combining the memory of both CPU and GPU. This is a feature that AMD offer which Nvidia do not – and one of the main reasons why their benchmarking results were above team greens.
All being said, this is actually a serious talking point when purchasing a new GPU – mainly because users that currently own an Intel CPU probably won’t get the performance gains shown on AMD’s slides.
However, with the market seemingly shifting in AMD’s favor, are many people that worried?
So, there you have it, our comprehensive look at the latest the RX 6000-series vs RTX 30-series GPU lineups. I have to say, I’m extremely impressed with how AMD has come back after the clear lack of interest shown by consumers in the late 2000s. As Dr. Su always says, AMD is on a journey, a journey that will leave them at the top of the GPU hierarchy. Whilst this may sound a little Holywood for most people, it’s clear that that’s the case – especially when you look at the products they’re bringing out in more recent times.