Ive compared Nvidias RTX 3080 desktopgraphics card against the 3080 laptopin 17 different games at 2 resolutionsto find out all of the differences. Here are the differences in specs betweenthe 3080 mobile and desktop GPUs. Both areGA104 dies, but the desktop 3080 almosthas 42% more CUDA cores compared to thelaptop version. The laptop GPU is availablewith either 8 or 16 gigs of GDDR6 memory,while the desktop GPU has 10 gigs of faster GDDR6Xmemory. Both are listed with the same boost clock,however the desktop cards powerlimit is more than twice as high.
The laptop 3080 has to have a lower GPU powerlimit because more power equals more heat,and I dont know if you can tell, butthe size difference between these twosystems is pretty big. Basically thelaptop just has less space for cooling. To do this testing Im using the XMG Neo 15 forthe laptop, its got an 8 core Ryzen 9 5900HXprocessor and 16 gigs of DDR4-3200 memory indual channel.
Now Ive chosen to use this modelbecause it both has an RTX 3080 GPU that runs atfull power limit but it also has a MUX switch,meaning we can bypass the integratedgraphics and get full performance.Something that we dont have toworry about with our desktop here.
Now although my particular laptop here runsthe RTX 3080 between 150 and 165 watts,it is worth mentioning that the 3080 can godown to 80 watts based on Nvidias spec sheet,and a lower GPU power limit on thelaptop side will mean lower performance.So consider this comparison as more of abest case. Not only does my 3080 have thefull power limit, but its also the 16gig version rather than the 8 gig one.
The desktop PC also has an 8 core Zen3 processor, but its the Ryzen 7 5800Xand like the GPU that does run at ahigher power limit compared to thelaptop. Ive also used 16 gigs ofDDR4-3200 dual channel memory here,granted the timings are lower, which isjust typically the case in a desktop system.
Its pretty important to note that theGPU power limit difference between thesetwo systems is about 170 watts. Just for arough comparison, when I previously compareda 3070 desktop against a 3070 laptop,we were looking at a 95 watt difference,and when I compared a 3060 desktop against a3060 laptop there was just a 55 watt difference.So basically Im just trying to say thatwe should expect a much bigger differencebetween the 3080 desktop and laptop becausethe power limits between them is even bigger.
Alright so with all of that in mind lets getstraight into the benchmarks and find out howthe desktop RTX 3080 compares against thelaptop 3080. Well start out with those 17different games at 2 different resolutions thenafterwards well compare things like power draw,content creator workloads and thenall important pricing & availability.
Lets start out with Far Cry 6, which was testedwith the games benchmark. Ive got the 1080presults below and the 1440p results above, withthe laptop results underneath the desktop results.Even the 1% lows from the desktop 3080 were higherthan the average FPS coming out of the laptop,while the average FPS was about 48% higher on thedesktop at both resolutions. This game launchedwith FSR support, and with this enabled thedesktop 3080 actually lowered back a littleat 1080p. The laptop 3080 increased a coupleof frames, but both saw larger differences atthe higher 1440p resolution, which Id argue ismore important for a feature like FSR anyway.
Cyberpunk 2077 was tested in little China onboth machines. Again the desktop had some prettylarge leads and was reaching a 40% higher averageframe rate at the lower 1080p resolution and a 50%higher average frame rate atthe higher 1440p resolution,granted the laptop wasnt far off60 FPS even at 1440p high settings,but the desktop was significantly better.
Thisgame has DLSS support, which both systems areable to take advantage of, and yeah with thelaptop now above 80 FPS high settings 1440p,its still running well, but theres noignoring the desktops 1% low being close by. Red Dead Redemption 2 was tested with the gamesbenchmark. At 1080p the desktop 3080 was reaching56% higher average FPS compared to thelaptop 3080, the third biggest difference atthis resolution out of all 17 games tested, butthen at 1440p the desktop was nearly 60% ahead,the biggest difference out of all gamestested.
This game also has DLSS support,and while the boost at 1440p is largercompared to the changes at 1080p,both GPUs are only receiving arelatively small 5 FPS boost,at least with DLSS set to quality. Different modeswill of course affect performance differently. Call of Duty Warzone was testedwith all settings maxed out. The 1%lows from the desktop 3080 were close to theaverage FPS of the laptop 3080 at 1080p, theneasily ahead at the higher 1440p resolution. Interms of average FPS though, the desktop was 36%faster at 1080p and 40% faster at 1440p, soundshigh but these are below average results comparedto the other 16 games tested.
This is yetanother game with DLSS support, and bothGPUs are able to get reasonable performanceimprovements with this one simple change. Control is a fairly GPU heavy game,which may explain why the differenceat 1080p was the biggest out of all17 games were looking at with a 58%higher average frame rate coming from the desktopsystem. There was still a well above average 55%lead by the desktop at 1440p. This gamehas a great implementation of DLSS,and the performance boosts are quite large.Yeah the desktop 3080 is still quite far aheadof the laptop, but above 100 FPS max settingsat 1440p on the laptop is plenty for this sortof game.
Ive also tested both with ray tracingenabled. Sure, DLSS could have also been usedhere to boost performance, which would berecommended given the frame rates at 1440p,but I wanted to compare pure RT differences.Theres a similar 60% boost at 1440p with thedesktop 3080 here, so a little more of a leadcompared to with RT off, though not by much. F1 2021 was also tested with some raytracing, because the maximum ultra highsetting preset enables some of these settingsby default, as listed at the top of the graph.This one also had above average gains forthe desktop 3080 with the games benchmark.
Microsoft Flight Simulator was testedin the Sydney landing challenge,and once more some fairly large gainswith the desktop, but again Id stillargue that at 1440p above 60 FPS on high settingsis quite good from a laptop and absolutely usable,but yeah if you want the best possibleexperience, clearly desktop is the way to go.
Watch Dogs Legion was also testedwith the games benchmark, but hadsome of the smallest differencesbetween the two test systems outof all games tested. At 1080p thedesktop was 27% ahead of the laptop,the smallest of all differences, and was then32% faster at the higher 1440p resolution,where the 1% low from the desktop was stillhigher than the laptops average frame rate. Assassins Creed Valhalla tested withthe games benchmark too, and this wasanother that had smaller differences, around 30%higher average FPS with the desktop here.
A 30%gain sounds decent standalone, but most othergames had even higher boosts on the desktop.
Instead of talking through the restof the 8 games individually in depth,Ill just blast through them so wecan get into the average differencesas thats more interesting, even if all ofthis testing did take many hours to complete.A larger selection of games is importantto help get a better rounded average.Feel free to pause the video if you want acloser look at any of these additional games. On average over all 17 games tested at 1080p, thedesktop RTX 3080 was 45% faster than the laptop,a larger difference compared to the 28%difference we saw when comparing a 3070 laptopand 3070 desktop.
Even the worst case resultsdown the bottom of the graph are near that,while in the best case were almostseeing some 60% gains in Control.
Stepping up to the higher 1440p resolutionthe performance difference is slightly biggeron average, with the desktop 3080 nowreaching 47% higher average frame rate.Again best case were looking at closeto a 60% improvement with the desktop,while worst case theres at least a 31% boost. The desktop system is definitely usingmore power when compared to the laptop.Both were measured at the wall, so this isthe total power that the entire machine isusing and not just the GPU. In control at4K, a GPU heavy game at a high resolution,the desktop system was drawing around 78% morepower from the wall, so a larger differencecompared to the performance boost on offer,meaning the laptop is a bit more efficient.
So when I compared a 3060 desktop against a3060 laptop, the desktop was just 9% faster,and then when I did the same with a 3070 desktopand a 3070 laptop the desktop was about 28%faster, and now when were comparing thetop end RTX 3080, the desktop was about 48%faster at 1440p. So were definitely seeingbigger differences with the higher tier GPUs,simply because they have bigger power limits. Considering the performance difference betweenthe 3080 desktop and laptop is quite a bit biggercompared to the 3060 and 3070 it does kind offeel a little weird to call the laptop a 3080.I mean I could understand if you think its a bitmisleading after you see a review on the desktop3080 and then you go to the shop and you find a3080 gaming laptop. I assume you probably thinkits probably got similar performance, and thatwould be fair enough based on Nvidias marketing.
I did actually ask Nvidia about this whenthey originally launched the 30 series and mytakeaway was that the 3080 laptop is basicallya state of mind, because basically what Nvidiatold me is that the 3080 on the laptop siderepresents the best case performance you canget on the laptop rather than it being directlycomparable to the desktop card with same name.
Its an experience bro, you wouldnt get it Just before we get into the most importantdiscussion of pricing and availability,lets check out some content creatorworkloads, its not all just about gaming. Blender was tested with CUDA, and the results werequite close together. The longer classroom testwas 27% faster on the desktop 3080, though as aCUDA workload I was expecting a bigger difference. The difference was a little larger in V-Ray,with the desktop 3080 scoring30% higher than the laptop 3080,so not exactly scaling with the 42% higherCUDA core count found with the desktop card.
SPECviewperf tests out various professional3D workloads. The desktop 3080, shown by thered bars, was always ahead ofthe laptop in the purple bars,but the margins could vary significantlydepending on the specific test. So the desktop RTX 3080 was clearlybetter compared to the laptop RTX 3080,surprising no one. The next question iswhats the price difference between these two?Pricing and availability willof course change over time,so refer to those links inthe description for updates.
Right now the price of a 3080 laptop canvary significantly based on the specificmodel. On Newegg theyre startingat $1900 USD with the MSI GP66,though its worth noting the GPU power limitis a little lower than what Ive tested with,plus it has half the VRAM at 8 gigs, grantedthat should still be plenty for most games. The exact same laptop I tested atEluktronics is more expensive at $2800for comparison, honestly probably not worththe massive 47% price increase over thatprevious GP66 example as the performancedifference is probably relatively low.
For a desktop PC, well, we all know thatgraphics card prices have been all over theplace for months now. The RTX 3080 was meantto have a $700 USD MSRP, talk about joke ofthe year.
Right now on Newegg the lowest one instock is more than double at close to $1600 USD. So just to be clear, right now the cheapestRTX 3080 desktop graphics card is about $300less compared to the cheapest RTX 3080laptop. A complete gaming system ready to go. The rest will greatly depend on the other partsthat you select, the 5800X I tested with at leastrecently lowered in price at Microcenter afterIntels 12th gen launch, so thats good newsat least, but that alone is $300, so CPU plusGPU alone is now the same as MSIs 3080 laptop.
Of course if you already have a desktopPC then simply taking out your existinggraphics card and slotting in a newone is a much easier upgrade path,and one thats not available to most laptops,but if youre buying a whole new system and youhave to factor in everything else like case,power supply, CPU, RAM, motherboard and morethen the desktop PC is quickly going toend up costing more than most laptops.
At least speaking generally, because thisparticular laptop is quite expensive comparedto most others out there. At the end of the day,to me personally the cost difference betweenthese two isnt even the most important factor. Doyou actually need to take your machine with you?If you do then good luck taking something likethis to school or the office.
For people thatdo actually need portability then clearly alaptop is better.
Or maybe its another reason,like you just dont have enough desk spaceto keep a full system like this laying aroundall the time. Laptops clearly take up lessspace and not everyone has a lot of room.And with all the recent supply issues, maybeits easier to find a laptop ready to go.Now thats not to say that laptops haventbeen subject to the same supply shortages,but generally speaking its abit easier to go out and findone whole machine compared to finding everysingle part that you might need at a good price.
It all really depends on your ownpersonal use case, but for most peopleI think if you do need portability then youneed a laptop.
If you dont need portabilitythen absolutely get a desktop. Obviouslythe desktop is way easier to upgrade,I cant change the CPU and GPU on this laptop, butnext time theres a new graphics card available Ican take out the 3080 and put in the newone – of course assuming its in stock.
Anyway I guess the point Im trying tomake is there are different solutions fordifferent people. I just get way too manycomments on these sorts of videos sayingbuying a laptops a mistake, justget a desktop, again it depends. I guess at least we do live in aworld where we do have these options.
Check out my comparisons between the laptop anddesktop RTX 3060 and RTX 3070 over here next.The difference between those lower tier graphicsoptions is actually smaller compared to what wesaw here due to lower power limits. The 3060 onein particular is very interesting, because thedesktop actually wasnt that much faster than thelaptop. Ill see you over in one of those next!.