Should you get AMD’s RX 6800M or Nvidia’s RTX 3070 graphics in your next gaming laptop? Let’s find out! These are the differences in specs between these two laptop GPUs. The RTX 3070 has 8 gigs of GDDR6 memory while the RX 6800M has 50% more at 12 gigs, but Nvidia’s memory has more bandwidth and a larger memory bus. CUDA cores and stream processors aren’t directly comparable, so it doesn’t mean much that the 3070 has twice the amount, though the 6800M has a higher maximum power limit.
To do this testing I’m using the Neo 15 from XMG for the RTX 3070 and the Strix G15 Advantage Edition from ASUS for the RX 6800M. Both laptops were tested with the exact same physical kit of memory, 16 gigs of DDR4-3200 CL22 x8 single rank memory in dual channel. Unfortunately there is a CPU difference at play here, as the Strix G15 has the Ryzen 9 5900HX while my Neo 15 has the Ryzen 7 5800H. Both are still 8 core Zen 3 processors and when I compared them in this video there wasn’t that much of a difference, and given we’ll be focusing on both games at higher setting levels and we’ll be testing 1440p as well I don’t think this will matter that much. I just couldn’t get both laptops with the same processor.
I have tested both laptops with an external screen though to bypass the integrated graphics and get best performance.
My 3070 laptop can boost up to 140 watts with Nvidia’s Dynamic Boost, but it will depend on the workload, and likewise the 6800M laptop also shifts power between the CPU and GPU with AMD’s SmartShift which works on a similar principle. Alright so let’s start out by comparing both laptops in 18 different games at both 1080p and 1440p resolutions, then afterwards we’ll compare things like pricing, content creator workloads and power draw. Let’s start out with Far Cry 6, which was tested with the game’s benchmark. I’ve got the 1080p results down the bottom and the 1440p results at the top, with the RX 6800M above the RTX 3070 results.
The 6800M was a little ahead when it came to average FPS at both resolutions, however the 3070 was giving us a higher 1% low.
This game launched with AMD’s FSR support, and this can be used to boost performance for both AMD and Nvidia GPUs. With FSR set to quality mode, the 1080p results hardly change on either GPU, though there’s more of a boost at 1440p, which I’d argue is where you’re more likely to use a feature like this on either of these high end GPUs anyway. FSR also moves the 1% low of the 6800M ahead of the 3070 at 1440p now. Far Cry 6 is an AMD sponsored title, so to keep things fair here’s how Control does on both GPUs as an Nvidia sponsored title.
Despite the fact that this game generally runs better on Nvidia hardware, both the 3070 and 6800M were reaching basically the same average frame rates, however the 3070 did have consistently higher 1% lows, so a more stable experience.
This game has DLSS support, and unlike FSR this is only available for Nvidia graphics. The performance boost that it’s able to offer is kind of crazy, even at 1440p the 3070 is above 100 FPS now, while its 1% lows are well above even the average frame rates coming out of the 6800M. This game supports ray tracing, and while I could have also combined DLSS on the 3070 to boost it even further, I just wanted to compare pure RT performance here. Neither are particularly great at 1440p, at least in terms of average FPS, but the 3070 was reaching 19% higher average FPS.
Cyberpunk 2077 was running better on the 6800M laptop. At 1080p it was reaching 7% higher average frame rate over the 3070, and then a slightly bigger lead with an 8% increase at the higher 1440p resolution.
This is another game with DLSS support though, so the 3070 is able to pull off a fairly nice win with this setting enabled. Interestingly the 1% low at 1080p was still better on the 6800M, but then at 1440p the 3070’s 1% low is now higher than even the 6800M’s average FPS. Red Dead Redemption 2 was tested with the game’s benchmark, and the 6800M had a fairly large 18% higher average frame rate at 1080p, granted at nearly 100 FPS it’s not as if the 3070 is unusable or anything.
The 6800M’s lead increases at the higher 1440p resolution and it’s now reaching 21% higher average FPS compared to the 3070, the biggest difference at this resolution out of all 18 games tested, at least until DLSS comes in. This game added DLSS support not too long ago, however in quality mode it wasn’t giving that much of a performance increase, at least when compared to the previous games that were tested with DLSS enabled. It’s not enough to beat the 6800M. I thought Watch Dogs Legion might be interesting as the game’s benchmark at max settings indicates that the 8 gigs of VRAM available to the 3070 were only just enough. Despite the 6800M having an additional 4 gigs of VRAM though, the performance difference was small between them.
At 1080p the 6800M was only 1 frame or so better in the averages, though there was a larger 8% boost to the 1% low. Perhaps the VRAM difference starts mattering more at 1440p, as the 6800M’s lead increases more, granted we’re only talking a 3 FPS difference here, but the 1% lows again had a much larger increase with the 6800M which were now 24% ahead of the 3070. This game also has DLSS support, and with this on quality mode the 3070 was able to pull out in front in terms of average FPS, and even the 1% low was able to get a subtle lead at 1440p, though it was still behind at 1080p. Things were much closer together in Microsoft Flight Simulator, at least at 1440p where both had pretty much the same average FPS just below the magic 60, however the 6800M had the edge when it came to 1% low in this one.
The gap was larger at the lower 1080p resolution, and although the 6800M was ahead, at just 3 FPS it’s nothing major.
I’d argue that the 12% boost to the 1% low with the 6800M is more important. Assassin’s Creed Valhalla had above average gains with the 6800M, which I suppose isn’t that surprising from an AMD sponsored game. The 3070 was still reaching above 60 FPS at the higher 1440p resolution with the second highest setting preset, so not bad at all, it’s just that the 6800M was 11% faster.
The gap was actually larger at the lower 1080p resolution. The 6800M had a 13% lead there, so it appears that Nvidia is able to get a bit more of a boost at higher resolutions, not always this case, but this is something we’ll see throughout many of the games covered.
Call of Duty Warzone was one of the best results for the 6800M, reaching a 21% higher average frame rate at 1080p, the second largest difference out of all 18 games tested at this resolution, and there was a larger 25% boost to the 1% low too, so a more consistent experience on the 6800M with less dips. It’s a similar deal at 1440p too, almost a 20 FPS higher average frame rate puts the 6800M around 19% ahead of the 3070. This is another game with DLSS support, and by turning this on the 3070 was able to get ahead of the 6800M in average frame rate, though it’s quite close, and the 6800M was still in the lead as far as the 1% low goes.
Although the 3070 still had a boost at the lower 1080p resolution, the 6800M still had the advantage. Fortnite is another eSports title, and one that generally favors Nvidia hardware.
The differences in average FPS at both resolutions were extremely small here, within a 1 FPS difference, margin of error stuff realistically. The 1% lows at 1440p were pretty close too, there’s a bigger difference to the 1% low at 1080p though, where the 3070 was 20% ahead of the 6800M. This game has DLSS support too, and enabling this is able to give a pretty big boost to the 3070. Even its 1% low was now ahead of even the average FPS from the 6800M at the higher 1440p resolution. That wasn’t quite the case at the lower 1080p resolution, however the 1% low was now 38% ahead on the 3070 compared to the 20% we saw prior to DLSS.
F1 2021 was tested with the highest setting preset which enables some ray tracing effects by default, as denoted at the top of the graph. There’s basically no difference between them at either resolution, at least not one you’re likely to notice when playing. I think this game is very close to adding FSR support, but unfortunately it wasn’t in the game when the testing was done. Instead of talking through the rest of the 8 games individually in depth, I’ll just blast through them so we can get into the average differences as that’s more interesting, even if all of this testing did take many hours to complete. A larger selection of games is important to help get a better rounded average.
Feel free to pause the video if you want a closer look at any of these additional games.
On average over all 18 games tested at 1080p, AMD’s Radeon RX 6800M was reaching 8% higher average FPS when compared to Nvidia’s GeForce RTX 3070. There are a few games towards the bottom of the graph that only saw minor differences one way or the other, including games like Fortnite and Control which typically seem to favor Nvidia. Best case in some games, we’re seeing a 21% boost with the 6800M, though as mentioned a small amount of this could be a result of the Ryzen 7 and Ryzen 9 CPU differences. If we step up to the higher 1440p resolution where we’re more GPU bound now, things shift over slightly towards the 3070, with the 6800M now 7% ahead when averaging all 18 games, a little less of a lead compared to the lower 1080p resolution.
I was wondering if the 6800M would pull out further ahead here with its 50% higher VRAM capacity, but that wasn’t the case, as the 8 gigs available to the 3070 seems to be enough in most modern games even at 1440p.
While generally performing better, it’s not without a cost, and I’m not talking money, but we’ll look at price difference soon. The RX 6800M laptop was drawing 25% more power from the wall when running Control at 4K, and as we saw performance was fairly similar in that specific game. I also want to note that I had no crashes or any other sorts of driver stability problems on either laptop, they both worked perfectly fine in all games. Alright so generally speaking, the RX 6800M was doing better than the RTX 3070 in games, but that said in some games the 3070 was pretty close or a little ahead of the 6800M.
I would have loved to include more games with FSR support but there still aren’t that many popular titles with it. Far Cry 6 is a good start though, but the fact is whether or not you like it, DLSS is currently supported in more games. As for the VRAM differences, the 8 gigs in the 3070 seems to be enough today even in modern games at 1440p with higher setting presets. Now it is absolutely possible that in future the extra 4 gigs in the 6800M might give it a further boost in newer games, but that’s really difficult to predict. Personally, I don’t think I would be making the purchasing decision based purely on the VRAM difference.
Maybe by the time there is a more practical difference between 8 and 12 gigs of VRAM you’ll be looking for a new laptop anyway. So then, what about price difference? Prices of both will of course change over time, so refer to those links in the description below for updates. At the time of recording, the ASUS Strix G15 Advantage Edition with the same specs that I’ve tested here goes for $1650 USD. This is quite well priced for the performance on offer, and as far as I’m aware it’s still the only gaming laptop with a 6800M available right now.
Generally speaking, 3070 laptops are more expensive than this, but the 3070 is available in a lot more machines, so it depends. Cheaper 3070 laptops might not have the full wattage power limit, so performance would be lower. The XMG Neo 15 I’ve tested is more than 2100 Euro, so considerably more. That said, I’ve seen other 3070 models like the Legion 5 Pro on sale with full powered GPUs for about $1500 USD. That’s a good sale price though, right now a Legion 5 with similar specs is more than $1900 right now.
Interestingly from a cost per frame perspective, with the sale price of that Legion 5 Pro it ends up offering about the same value as the more expensive RX 6800M machine. The 6800M laptop costs $120 or 8% more money, but it’s performing about that much better too, which is why the cost per frame values are so close. Without the sale though, a full priced 3070 Legion 5 right now is significantly worse value in terms of dollar per frame when compared to the 6800M. Basically this means unless you’re getting a full powered 3070 for $1550 USD or less, the RX 6800M for $1650 is just better value. At least in terms of gaming performance and that’s not factoring in things like FSR or DLSS.
Of course there’s way more to buying a laptop that just the GPU that’s inside. You can check out my full review videos on both of these laptops over here next if you want to find out both the good and the bad.
Make sure you’re subscribed for future laptop comparisons like this one, and come and join and the community in Discord and get behind the scenes videos by supporting the channel on Patreon..
Read More: 12700K vs 11700K vs 10700K vs 9700K vs 8700K Comparing 5 Generations of Intel i7 Processors!
