How much better is Nvidias desktop RTX3070 graphics card compared to the laptop3070? Ive compared both in 17 differentgames at 1080p and 1440p to find out! Here are the differences in specs between the3070 mobile and desktop GPUs. Both are GA104 dies,but the desktop 3070 has 15% more CUDAcores compared to the laptop version.They both have 8 gigs of GDDR6 memory at thesame speed, but the desktop card is of courseable to hit higher clock speeds asits got a higher 220 watt power limit. The laptop 3070 has to have lower power limitsbecause more power generally equals more heat,and I dont know if you can see the sizedifference here but its pretty big!Basically the laptop justhas less space for cooling. To do this testing Im usingthe XMG Neo 15 for the laptop,its got an 8 core Ryzen 7 5800H processorand 16 gigs of DDR4-3200 memory in dualchannel.
Ive chosen this model becauseit both runs the RTX 3070 mobile at itsfull power limit and it has a MUX switchallowing us to disable optimus and bypassthe integrated graphics – something thatour desktop system here isnt subject to. Now technically the Nvidia spec sheet doesallow the RTX 3070 mobile to run all theway down to 80 watts, so expect an evenbigger difference compared to what Illbe showing with these two here.
Thisis basically a best case 3070 laptop. The desktop PC also has an 8 core Zen3 processor, but its the Ryzen 7 5800Xand like the GPU that does run at a higherpower limit compared to the laptop. Ivealso used 16 gigs of DDR4-3200 dual channelmemory here, granted the timings are lower. Alright lets get straight intocomparing laptop 3070 againstdesktop 3070 in 17 different games at 2different resolutions, then after thatwell compare pricing and availability,power draw and content creator workloads. Lets start out with Red Dead Redemption 2, whichwas tested with the games benchmark.
Ive got the1080p results below and the 1440p resultsabove, with the laptop results underneaththe desktop results. At 1080p the desktop 3070was reaching 18% higher average FPS comparedto the laptop 3070, a below average result outof the 17 tested, with a similar 19% lead at1440p. This game recently added DLSS support,and with this enabled both were able to get afurther performance boost. 85 FPS on a laptop athigh settings 1440p is by no means unplayable,so although the desktop is still around 19%higher, the laptop still offers a good result. Cyberpunk 2077 also had a below average differenceout of all of the games tested.
High settingsat 1440p was just above 60 FPS on the desktop,putting it 24% ahead of the laptop. Even the 1%low from the desktop was ahead of the average FPSfrom the laptop. The difference was similar at thelower 1080p resolution with the desktop systemalmost 26% faster than the laptop now. This isanother game with DLSS support, and both machinesare able to get a decent performance boost byenabling this simple setting, though the desktopwas still 26% ahead of the laptop at 1080p,so no change there, but the laptop was now capableof averaging above 100 FPS at high settings. Call of Duty Warzone was tested withall settings maxed out.
The 1% lows fromthe desktop 3070 were close to the averageframe rates that laptop was able to offer,though with the laptop capable ofabout 100 FPS at the higher 1440presolution at max settings its not exactlyterrible, despite the desktop reaching 30%higher average FPS.
At the risk of sounding likean Nvidia shill, this game has DLSS support too,and both machines can get an FPS boost by enablingthis setting. The desktop 3070 was now about 28%ahead of the laptop, so a similar margin,and both were running the game without issue. Control is a fairly GPU heavy game, though thedifferences between the two machines were closeto the 17 game average. The desktop had 1% lowshigher than even the average FPS from the laptop,a common trend in this testing.
Again Id argue60 FPS at max settings 1440p on a laptop is quitedecent, but the desktop 3070 was reachinga 32% higher average frame rate compared tothe laptop here, and had a similar 31% leadat 1080p. This game also has DLSS support,so yeah now that were above 100 FPS at 1440pon the laptop theres really no concern aboutplayability or anything, but the fact isthe desktop was now 34% ahead of the laptop,so its receiving a larger gain. Ive alsotested both with ray tracing enabled.
Sure,DLSS could have been used here to boostperformance, which would be recommendedgiven the frame rates at 1440p, but I wanted tosee the RT difference between laptop and desktop.Both resolutions were around 27% fasteron the desktop, a lower difference.
F1 2021 was also tested with some ray tracing,because the maximum ultra high setting presetenables some of these settings by default.Again 1% lows from the desktop were beatingthe average FPS from the laptop. The desktopwas reaching 33% higher average FPSat 1080p and 32% higher at 1440p. Microsoft Flight Simulator wasright in line with the 17 gameaverage at 1080p with the desktop 3070reaching 30% higher average frame rate,though in this game I wouldnt be surprisedif processor difference was playing a part,but thats just the way it is when comparinglaptop and desktop.
At 1440p the desktop wasreaching 33% higher average FPS compared to thelaptop, the 4th best result at this resolution.
Watch Dogs Legion on the other handhad one of the smaller differencesout of all games tested, atleast in terms of average FPS,the 1% low gap seems to be larger. At 1440p thedesktop was reaching 24% higher average FPS,but the 1% low is almost 49% higher, so amuch more stable experience on the desktop,but again its hard to say how much of this isdue to other components such as the processor,its not just the GPU that has a higherpower limit in the desktop, the CPU does too.This game also has DLSS support which can beenabled to further boost performance on both GPUs. Shadow of the Tomb Raider hadthe largest difference out ofall 17 games tested in favorof the desktop 3070. At 1080pthe desktop had a massive 47% lead over thelaptop, and then it was 36% ahead at 1440p.
Assassins Creed Valhalla on the other handhad the smallest difference between the laptopand desktop out of the 17 games tested.
Thedesktop was reaching just 11% higher averageFPS at 1080p and 15% higher at 1440p. Giventhe laptop is still above 60 FPS here though,like most of the other games covered,its not like its going to be unplayable,and thats coming from someone who hasplayed this game for more than 100 hours now. Fortnite was tested with the exact same replayfile on both machines. Like many other games,the 1% lows from the desktop were aheadof even the average FPS from the laptop,but 100 FPS at max settings 1440p onthe laptop is of course still playable,the desktop is just reaching33% higher average FPS.
Instead of talking through the rest of the7 games in detail, Ill just quickly skipthrough them so we can get onto looking atthe average differences of all games testedas thats more interesting, just pausethe video if you want a detailed lookat any of these results.
Differencescan vary a fair bit between games,which is why its important to test a largesample so that we can get a better average. On average over all 17 games tested at 1080p, thedesktop RTX 3070 was 28% faster than the laptop,much more of a difference compared to myrecent comparison between the laptop anddesktop 3060 which only saw the desktop 8%faster than the laptop in these same games.I suppose the 3060 laptop and desktop have a lower55 watt power difference, while the 3070 laptopand desktops gap is larger at 95 watts, and Ibelieve this is why the difference is larger here.
Stepping up to the higher 1440p resolution theperformance difference is slightly smaller onaverage, but still quite similar and close tothe 29% point.
Most of the games covered seemto be around 30% faster on the desktop, butyeah as you can see it depends on the game,worst case Valhalla down the bottom wasless than 15% faster on the desktop. The desktop system is definitely using more powerwhen compared to the laptop. Both were measured atthe wall, so this is the total power that theentire machine is using and not just the GPU.In control at 4K, a GPU heavy game at ahigh resolution, the desktop system wasdrawing around 65% more power from the wall,and granted I didnt test FPS at 4K, 1080p and1440p resolutions were about 30% higher averageFPS on the 3070 desktop compared to the laptop,so theres definitely an argument to be hadfor the laptop being more power efficient. Just before we get into the price differences,Ive also compared both in some content creatorworkloads, its not all just about gaming.
Blender was tested with CUDA, and theresults were quite close together.
Thelonger classroom test was nearly 12% faster onthe desktop 3070, which I suppose makes sensewhen you consider that the desktop 3070 has15% more CUDA cores and this is a CUDA test.It would seem that the power limit differencemeans less here compared to the games previously. Likewise V-Ray was also tested with CUDA,and the desktop card was scoring 18%higher than the laptop graphics here.Ive also run its ray tracing test,and the desktop was scoring about9% higher than the laptop there.
SPECviewperf tests out various professional3D workloads.
The desktop 3070, 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 3070 was generallybetter, surprising absolutely no one.The next question is whats the pricedifference? Prices will of course changeover time so refer to those links inthe description below for updates.
Right now the price of a 3070 laptop canvary significantly based on the specificmodel.
On Newegg were generallylooking at $1800 to $2200 USD.The XMG Neo 15 Ive tested is currently 2200 Euro,but Ive seen other full powered 3070 laptopslike the Legion 7 on sale for about $1600 USD. For a desktop PC, well, we all knowthat graphics card prices have beenall over the place for months now. The RTX3070 was meant to have a $500 USD MSRP,but we can see on Newegg that the cheapestone currently in stock is $1050 USD,with most being more than this, even thelowest out of stock ones are mid 700s.
The rest will greatly depend onthe other parts that you select,the 5800X Ive used for example is another $400alone, but honestly something like a 5600Xfor less money will probably still do well. Of course if you already have a PC,just upgrading the GPU is also an optionthats not available with a laptop. Im not going to go ahead and do a complete PCbuild because as mentioned prices are just allover the place, and its going to heavilydepend on what components you pick anyway,but once you factor in everything likemotherboard, case, power supply, RAM, storage,screen, keyboard, youre probably going to end uppaying a similar amount to a laptop anyway.
Butas mentioned it really will depend on the hardwareselection, again Ive got absolutely no doubt thatyou can find cheaper desktop components and even acheaper 3070 laptop, Im just speaking generally. At the end of the day, to me personallycost isnt even the major factor.
Itcomes down to whether or not you actually needportability. Do you actually need to take yourmachine with you? If you do, good luck takingsomething like this to school or the office.For people that do need portability, a laptopis clearly the superior option, or it might evenbe another reason like just not having enoughdesk space to keep something like this setup,or maybe as this short foretold youre havinga hard time getting your hands on these parts.
For some people it might be easier justto find a laptop in stock.
It depends,the point is obviously there are scenarios wherea laptop has its benefits. Generally though if youdo have the space for a desktop PC and you dontneed portability I would suggest a desktop PC. Mostly because its way easier to upgradecompared to a laptop. With a laptop the CPUand GPU are soldered to the motherboard and youcant upgrade them unless you replace the entiremachine.
Kind of inefficient, whereas onthe desktop I can just pull out this 3070and put the next gen graphics card straightin, of course assuming theyre in stock.
Anyway I guess the point Im trying to make isdifferent things for different people, and whilethe desktop 3070 was a fair bit better comparedto the laptop 3070, the fact is the laptop3070 was still offering a pretty great gamingexperience even at the higher 1440p resolution. Personally, what I was most interested in wasto see the performance gap between the laptop3070 and desktop 3070 being larger comparedto the laptop and desktop 3060, and thatsbecause the power limit difference between thesetwo is higher compared to the 3060 versions.Now the power limit difference between the 3080laptop and 3080 desktop is even bigger than this,so let me know if I should revisit thisvideo but using the RTX 3080 instead.
You can check out my comparison betweenthe laptop and desktop 3060 over here,and Ive also done the same typeof comparison with previous gengraphics over here, which might be easierto get your hands on than this new stuff.Make sure youre subscribed for futurecomparisons like this one, and come andjoin me in Discord and get behind the scenesvideos by supporting the channel on Patreon.
Read More: Review Alienware m15 R6 – Too Many Problems!