XMG Neo 17 RTX 3080 Laptop Performance Review vs Asus Zephryus G15 vs Desktop Ampere!
This is a gaming laptop that contains an rtx 3080 and a horizon 5000 octa core cpu nvme storage and a 1440p high refresh rate display quite the package. Moving from a desktop machine to a laptop. There are pros and cons, but it’s important to point out that the rtx 3080 in this thing isn’t really the same thing as this. So the laptop in question, the xmg neo17 it’s a beast. The focus of this review is going to be on the 3080, though. Why well, first of all, xmg had very limited review samples and the only one i could source didn’t have a working display. So i don’t think i can review this as a full laptop experience. I can only really tell you what it’s capable of behind the display the power of the silicon, if you like. Secondly, well, we really need to talk about gpu naming conventions. The 3080 in the neo 17 is extremely capable, it’ll likely outperform the vast majority of desktop gpus. It is the fastest laptop gpu it can do. Raid facing dlss is just as much of an accelerant here as it is on the desktop, but it is a 3080 that isn’t quite a 3080 here’s the specs table and it’s pretty clear that the laptop version of the 3080 has much more in common with the Desktop 3070 it’s, based on the same core, ga104 silicon with the same 256, bit memory interface, hooked up to gddr6 memory, not g6x.
Curiously, the laptop 3080 is actually more capable in one dimension than the 3070.. It has more cuda cores 6144 up against 5888. So this is actually a fully enabled ga104 chip, something that isn’t available in the desktop space. Certainly not yet. However, you can see that the desktop part is pumping a ton more power into the silicon 220 watts up against a maximum 150 watts on the mobile 3080. So the mobile part process is a little wider with more shaders, but has to do so with less power, but comparisons against the rtx 3080 desktop part. Do the laptop, no favors really there’s a massive deficit in terms of shaders and memory, bandwidth uh? Well, the whole business really so can the rtx 3080 laptop part really be called a 3080 it’s complicated. The 3080 is synonymous with best in class, desktop performance and the laptop 3080 does the same thing in the mobile space. So maybe it should be called something like a 3080m well yeah, possibly but yeah. It gets even more complicated than that. You see. The mobile 3080 might only be pushing around half of the wall power through its silicon and it might well be running with big deficits in memory. Bandwidth and clock speeds, but performance doesn’t scale perfectly in line with any of these things. So let’s dig into some performance differentials i’m using the xmg neo here with power configured to 135 watts. However, you also have something called dynamic boost.
If the cpu doesn’t need all of its power, an extra 15 watts can be diverted to the gpu. So in these benchmarks that we’re going to be looking at here, effectively we’re, looking at 150 watts total. What you’re looking at here is borderlands 3, where i’ve got the top spec rtx 3080 mobile running against desktop versions of the 3080 3070 and 3060.. I have a ryzen 9 5000 on the laptop here, but i’ve got a core i9 10 900k on the desktop side, but i think it’s pretty clear that we’re entirely gpu limited here at 1440p on all systems. I chose 1440p as the output resolution for a couple of reasons, but mainly because 1440p, with high refresh rates is the new sweet spot for high end laptops paired with this class of gpu power. End result well expressed as percentage differentials. The desktop 3080 is 51 percent. More performant than the mobile equivalent in top spec form. However, the desktop 3070 is only about 11 and a half percent faster, perhaps that’s, to be expected. Bearing in mind that the spec is so similar, but we’re running with a ton, less power, remember 220 watts on the desktop versus 150 on the laptop. The desktop 3060 has 73 of the performance level of the laptop 3080 or, to put it another way, the laptop 3080 is 37 faster than the desktop 3060. remedies. Control actually sees a wider performance. Differential 3080 desktop 62 percent faster than the laptop equivalent, while the full fat desktop.
30. 70. 21 faster, the 3060. Well, our laptop still beats it pretty comprehensively, but its lead drops to 27 points rather than the 37 we saw with boardlands, so doesn’t look so hot for the laptop right. Well, death stranding turns the tables. The desktop 3080 is only 35 percent faster than the laptop version, and now the 3070 only averages a 10 point lead over the mobile 3080.. I took a look at a bunch of titles and we can see some pretty wild fluctuations in the performance differentials. Generally speaking, our desktop 3080 is 40 to 60 percent faster than the laptop equivalent, but to get that performance increase, it requires over twice as much power. A 42 increase to shader count and a 70 boost to memory. Bandwidth, yeah that’s, pretty extreme. So can the rtx, 3080 and mobile form be equated to any specific desktop part? Well, you know it’s super super tricky, because the same silicone can be deployed in any given laptop with a different power budget. If you pop into the nvidia control panel, you should be able to get it to tell you what the tgp is. The total graphics power, but really this is information you need to know before you buy so let’s. Consider the neo17 here. What i did was to stack up the laptop against all major nvidia gpus in the 20 80 to 3070 window returning to borderlands 3. The laptop is overall 1 ahead of a 30 60 ti and 7 ahead of the 2080 super we’re talking about desktop parts.
There remember you may remember that control didn’t fare so well for the laptop in our last round of tests, and it is slightly under par here. 3060 ti is about four percent ahead. 2080 super three points to the better and again there’s, a very slight advantage for 3060 ti and 2080 super in shadow of the tomb raider. But you know i can cycle through a bunch of games and that’s, really the ballpark that we’re operating in i find the 3060 ti comparison quite fascinating, actually, because this gpu is based on the same ga104 processor as the mobile rtx. 3080 same memory bandwidth, but only 4 864 shaders up against the full complement of 6144 in a laptop part. However, the desktop part has a 200 watt power budget up against the 150 watts i’m, giving the mobile 3080. so doing some back of an envelope. Maths here 3060 ti only has 79 of the compute hardware, but does get equivalent or slightly better performance, because it has 25 percent extra power flowing through it. So i mentioned earlier that the xmg neo17 gives users the chance to scale power to configure how much juice the gpu gets. You can start at 115 watts with no dynamic power boost to 135 watts, with dynamic boost so effectively 150, so it’s quite a window there. Why would you reduce power well, to put it simply, you can trade performance for less heat, less heat means slower fan speeds, meaning less noise.
Now, as i mentioned earlier, there’s no linear relationship between power and performance, it’s more of a curve, that’s kind of borne out when you look at this i’m using metfor exodus here as an example, but i noted much the same results on all tested games. You can see how things scale in percentage terms here. The top three results are 135 125 and 115 watt tgp with dynamic boost enabled the bottom result is 115 watts without dynamic boost comparing the 3080 top and bottom. Basically, 77 of the power is delivering 88 of the performance, it’s slower, but it’s more efficient, easier to call quieter. I certainly noticed a big reduction in noise with the xmg here that increased efficiency also means that you could potentially fit the chip into smaller laptops and that’s. Exactly what’s happened, but i’ll have more on that in a bit but yeah increased efficiency is the order of the day. Do maternal again. 77 of the power is delivering an average of 88 of the performance, comparing 115 watts to 150. in death stranding, which has kind of weird scalability on nvidia hardware, where we see even better improvements in efficiency. Here 77 of the power is actually delivering 92 of the performance on the flip side, as we’ve seen with control it’s, not particularly friendly to our mobile rtx 3080, but the bottom line here is that 115 watts remember 77 of 150 watts max here. It averages out at 86.
4 of the overall performance, so slightly less efficient, but still not bad. So this is all really tricky to explain really in the laptop space. The product you’re being offered is about more than just performance, it’s about form factor acoustics. Dare i say it the experience, but if you are zeroing in entirely on performance, you do need to go into this with your eyes open, so i’ve not had the chance to go hands on with the asus zephyrus g15, but my hashtag friend and colleague will judd. Has and he’s shared the fame time data with me, and i can integrate it with my own results here to show you what’s going on so the asus it’s, a very different beast to the xmg it’s designed to be thin and light. The xmg is bigger. Bulkier designed for performance to achieve its stated design aims. Power targets for the 3080 are cut back pretty brutally on the asus. Now we can max out at 150 on the xmg, but we can reduce it to a minimum 115.. However, the asus it actually ships the same silicon, but with just an 80 watt tgp, but you can have an optional turbo mode. I’Ve got a couple of results for you here. Borderlands 3 fares pretty well despite the power differential turbo mode on the asus, gets close to the low end of the xmg there’s, a bit more of a gap in shadow of the tomb raider. As you can see here, but bearing in mind the power differentials, 80 watts on one end versus 150 – watts on the other, you’re still getting the majority of the performance.
So if the 150 watt rtx 3080 mobile is equivalent by enlarged to a 3060 ti desktop. Where does that leave a laptop part fed with a poultry 80 watts of power? Well, looking at borderlands 3 here, the asus beats the 3060, while the 95 watt turbo mode gives you point for point performance that matches an rtx 2080. Looking at shadow of the tomb raider, we moved down a notch. The 95w 3080 behaves a bit more like a 2070 super that’s, the closest desktop equivalent i could find by the way, but regardless fascinating stuff. So thus far, i seem to have spent quite a lot of time explaining how one gpu called an rtx 3080 isn’t, actually much like another gpu called an rtx 3080. While performance numbers see this new laptop 3080 performing like a desktop 2080 super or 3060 ti or a 2070 super or 3060 i’m kind of thinking at the least, the rtx 3080 mobile chip should be called a 3080m. So there is at least some kind of recognition that it’s not the same thing as the desktop version. The m designation was a thing back in the day, but it was dropped when 10 series came along and to be fair, those were the exact same chips as the desktop versions, with one or two exceptions, and the performance was close. Actually, things have moved on, though. The titanic efficiency of pascal isn’t quite the same with turing or the new ampere and the m designation.
I think it now makes a lot more sense, and maybe things are getting too complex here, but perhaps the xmg should ship with an rtx 3080m 150. While the zephyrus has a 3080m 95. kinda makes a bit more sense, then i think truth is. There are so many variables that can potentially affect performance here that it’s difficult to nail it all down, but ultimately i do feel that the confusion here can be at least mitigated. Things could be clearer bottom line, though the xmg neo17 i liked it, or rather i liked its performance, since i couldn’t really judge it as a laptop, owing to the fact that the screen wasn’t working on my sample now, this clearly isn’t a premium chassis design, but It’S all about wall power and that’s, where i think it’s a genuinely good, possibly even class, leading product. That performance is really impressive and quantifying that power. Well, the desktop 3080 is considered to be a 4k gpu right. Well. This opens up an interesting point of comparison because 1440p is the sweet spot in the latest range of laptops and the neo17, even at 115 watt tgp on its 3080 manages to beat a desktop 3080 running at 4k and the acoustics the fan, noise and whatnot it’s. Not bad when you throttle back the power, so yes, a laptop can provide a desktop like experience. High end, one at that i’d say maybe not the best of the best but the best, but when you’re on par or indeed beating rtx 2080 and 2080 super in their desktop incarnations when you’re retaining the entire feature set with raid facing dlss and all of that Jazz pretty good going in my opinion, that’s all for me right now, so yeah liking sharing subscribing.