One of the biggest new additions to Radeon 6000 series cards is Smart Access Memory (SAM), which is AMD's branded implementation of the PCI Express Resizeable BAR feature baked into recent builds of Windows 10 and Linux. SAM allows supported processors to directly access all 16GB of video memory on the Radeon 6800 or 6800 XT, bypassing the usual 256MB I/O buffer limit and thereby increasing performance in some games. AMD showed frame-rate gains in the region of two to 10 per cent in their Radeon 6000 announcement event, depending on the game and resolution, so let's put those claims to the test - and explain what you'll need to use the feature.
Right now, Smart Access Memory requires a particular setup: a 500-series motherboard (that's B550 or X570 at present) and a Ryzen 5000 processor (5600X, 5800X, 5900X or 5950X). AMD told us that there's no technical reason preventing SAM from working on other motherboards in the future, so perhaps we'll see the feature arrive on other motherboards down the line if there's enough interest from B550 and X470 system owners. Here's AMD's full statement:
"Currently, AMD Smart Access Memory is a feature for systems with AMD Ryzen 5000 Series processors, AMD B550 and X570 motherboards and Radeon RX 6000 Series graphics cards. There are no technical limitations to support SAM on other platforms, other than supporting certain features that are part of PCIe and firmware standards. SAM leverages the AMD device's resizable BAR feature and has been validated. We also tuned the driver to deliver optimal game performance with Ryzen 5000 Series processors and Radeon RX 6000 Series graphics cards. We're excited to show how it will improve PC gaming."
As well as having a supported motherboard, you'll also need to be using a motherboard BIOS released November 18th 2020 or later and a recent version of Windows 10 (our system is on 2004).
Once you've ensured you meet the requirements, you'll need to head over to your BIOS. (I like to get there by holding Shift while clicking Restart in Windows, then selecting Troubleshoot > UEFI Firmware Settings, but hammering Delete or F2 after a regular restart works too.) From here, each board is different, but on our Asus board we needed to visit the 'Advanced' tab, enter 'PCI Subsystem Settings', then enable 'Above 4G Decoding'. This unlocks a new option, 'Re-Size BAR Support', which when set to Auto with a compatible graphics card will activate SAM.
Interestingly, there doesn't appear to be a way to verify the setting is enabled in Windows - perhaps an indicator could be added to the Radeon control panel in a future update.
You may have noticed that the test rig in the rest of this review is an Intel system, so we needed to shift to AMD to test Smart Access Memory. For this, we used a Ryzen 9 5900X processor mounted in an Asus ROG Crosshair 8 Hero Wi-Fi motherboard using BIOS version 2402, released a few days prior. The rest of our setup remained the same, including an Alphacool Eisbaer Aurora 240mm AiO liquid cooler, 16GB of 3600MHz CL16 DDR4 RAM, a 2TB Samsung 970 Evo Plus NVMe SSD and an 850W power supply.
For each game we've included tests at multiple resolutions with SAM enabled and disabled, as well as a 4K 'baseline' result from our earlier testing with our Intel Core i9 10900K and Z490 motherboard test bed, which limits our Radeon 6000 graphics cards to PCIe 3.0 speeds.
Let's start with a game where AMD could really use some extra firepower: Doom Eternal. The RX 6800 and RX 6800 XT heavily outperform the last-gen Radeon models - we see a 90 per cent boost from the RX 5700 XT to RX 6800 XT - but the RTX 3080 still holds an eight per cent lead over AMD's best card to be released so far. Can Smart Access Memory be a difference-maker?
In short... no. At all three of the resolutions we tested, the RX 6800 XT doesn't see any improvements to frame-rate with SAM enabled, with each result falling well within run-to-run variance. Comparing our 10900K result, the 5900X manages 1fps more on average at 4K, but again this is well within the margin of error. Let's see if other games perform differently.
Doom Eternal: Vulkan, Ultra Nightmare, 8x TSSAA
Metro comes next; another game where AMD could use a small performance boost to counter a better-performing RTX equivalent. Here the Core i9 holds around 1 fps lead over the Ryzen 5900X at 4K, showing there's no penalty for opting for Intel's system despite the PCIe 3.0 limitation. There's also even less of a difference when enabling SAM, with all four runs coming within two tenths of a percentage point, including two identical runs with RT enabled.
Metro Exodus: DX12, Ultra, TAA
Assassin's Creed Unity
Assassin's Creed Unity is a game where the RX 6800 XT struggles, with the RTX 3080 scoring 12 per cent higher in our cutscene benchmark. It's also where we start to see our first signs of life for Smart Access Memory; turning on the feature leads to a two per cent boost at all three of the resolutions we tested. That's not much, but it's something! Our Intel system is again within one per cent at 4K, but the gap expands to around three per cent when you factor in Smart Access Memory.
AC Unity: DX11, High, FXAA
Borderlands 3 is a game where the RX 6800 XT already does well, and it was highlighted in AMD's reviewer's guide as being a title that really benefits from Smart Access Memory - so let's give it a try. We see very little difference between our Intel and AMD test systems at 4K, but turn on Smart Access Memory and the AMD platform gets out to a four per cent lead - not bad! This lead lengthens at 1440p to eight per cent, then again at 1080p to a mighty 12 per cent. That's a really respectable return, pushing the game at its highest settings from one that can max out a 144Hz monitor to one that can do the same with a 165Hz monitor. Note however that our lowest one per cent scores actually worsen slightly with SAM engaged, something that persisted after several retests, so that's something to keep an eye on.
Borderlands 3: Bad Ass, DX12, TAA
Now let's look at a game where Radeon graphics cards seem to absolutely struggle: Control. We recorded an average frame-rate of just 47fps at 4K, with the game running at its highest settings but all RT effects disabled. The RTX 3080 managed 58fps, a 23 per cent advantage, so a boost from SAM would work wonders here. And indeed, we do see solid improvement - just two per cent at 4K, but four per cent at 1440p and six per cent at 1080p. That's enough to close the 1080p gap between the 3080 and the 6800 XT from 12 to six per cent - not a complete fix, but a definite improvement.
Control: DX12, High, TAA
Our final SAM test reprises Mr. Alexander Battaglia's 'Corridor of Doom' in Control with all RTX effects engaged. The 3080 holds a massive 64 per cent advantage over the RX 6800 XT at 1440p, so how does SAM shift the balance of power? Unfortunately, not much at all - we spied just a one per cent margin for SAM enabled at 4K, and a two per cent boost from enabling the feature at 1440p. If we were seeing five to 12 per cent boosts consistently, then that would really go a long way to shoring up the RX 6800 XT's weaker RT performance, but unfortunately that just isn't the case right now. Still, it's better than nothing, and from our testing it seems that keeping SAM enabled is the way to go.
Control: DX12, High, High RT, TAA
So there you have it - some big gains in Borderlands 3 and non-RT Control, but unfortunately we don't see significant improvements in the other titles we tested that could shift the RX 6800 XT into closer contention with the RTX 3080. And contrary to our expectations, it looks like it's actually the lower resolutions, 1080p and 1440p, where the technology seems to be most effective, although the extra horsepower is needed more at 4K. Still, for the cost of enabling a BIOS setting, any performance gains are welcome.
It'll be fascinating to see a wider gamut of Radeon 6000 reviews to see which games really benefit from this feature - as if your go-to game sees a big performance uplift, going for an all-AMD system may make a lot of sense.
Of course, Nvidia is also working on their own implementation of the Resizeable BAR feature, so we'll have to see how long that takes to come to market - and if it offers the same performance benefits on Nvidia's Ampere architecture. Nvidia told us to expect similar performance, but didn't provide a timeline of when they expect to offer the feature on their own cards, suggesting it could still be some way out:
"The capability for resizable / larger BAR is part of the PCI Express specification. NVIDIA hardware supports this functionality and will enable it on Ampere GPUs through future software updates. We have it working internally and are seeing similar performance results. Stay tuned."
For now though, AMD's Smart Access Memory is well worth experimenting with if you're on a supported system and provides a unique - if perhaps shortlived - advantage for committed Red Team fans with the latest AMD processing and graphics hardware.
With that, we've reached the end of our game testing, so all that's left now is our conclusion. Let's wrap things up, as we consider how Big Navi has landed.
AMD Radeon RX 6800 and RX 6800 XT Analysis
- Introduction, Hardware and Power Analysis
- Doom Eternal, Control, Borderlands 3, Shadow of the Tomb Raider - Game Benchmarks Part 1
- Death Stranding, Far Cry 5, Hitman 2, Assassin's Creed Odyssey - Game Benchmarks Part 2
- Metro Exodus, Dirt Rally 2, Assassin's Creed Unity - Game Benchmarks Part 3
- Control, Metro Exodus, Battlefield 5 - RTX Game Benchmarks
- Smart Access Memory benchmarks and requirements [This Page]
- RX 6800 and RX 6800 XT - the Digital Foundry verdict
Digital Foundry specialises in technical analysis of gaming hardware and software, using state-of-the-art capture systems and bespoke software to show you how well games and hardware run, visualising precisely what they're capable of. In order to show you what 4K gaming actually looks like we needed to build our own platform to supply high quality 4K video for offline viewing. So we did.
Our videos are multi-gigabyte files and we've chosen a high quality provider to ensure fast downloads. However, that bandwidth isn't free and so we charge a small monthly subscription fee of £4.50. We think it's a small price to pay for unlimited access to top-tier quality encodes of our content. Thank you.Support Digital Foundry