RTX 3060 Laptop vs Desktop
NVIDIA’s ‘60’ class cards have always been the best sellers, both in desktop and laptop configurations. That makes sense, it’s generally the best bang for your buck you can get in the range while still getting good performance. But, how much more performance do you get from a desktop card over a laptop? We saw with the RTX 3080 that it’s between 30 and 50% slower than it’s desktop counterpart – but is this 3060 any better? Let’s test it and find out!
The first thing we need to establish is a comparative spec. The laptop I’m using there is the Acer Helios 300, which is running an RTX 3060, specifically with “up to 105W” listed as it’s TDP. I suspect this means it’s a 90W chip with up to 15W of dynamic boost available – and I’ll be showing results from both its standard and turbo modes.
The RTX 3060 Laptop chip in this has 3840 CUDA cores, which surprisingly is actually 7% MORE cores than you get in the desktop chip. It’s based on the same GA106 die, has the same 192 bit memory bus, although only has 6GB of VRAM compared to the desktop card’s 12GB. That does mean the desktop card is going to have a longer useable lifespan and would be suitable for 1440p gaming too. The other key difference is the TDP and clock speeds. The desktop card can boost up to nearly 1800MHz at stock and factory overclocked cards are closing in on 1900MHz, boasting a 170W TDP at stock. The laptop version? The highest TDP version you can get is 115W with 15W of dynamic boost, and it’s rated to boost up to 1425MHz.
As for the CPU, the Helios has an i7-10750H and 16GB of dual channel 3200MHz memory. The closest desktop CPU I have to that is an i5-10400F – it’s almost identical in it’s Cinebench scores both for single and multi threaded so this will work perfectly. I’ve also got 16GB of dual channel 3200MHz RAM for good measure.
So, the results! Starting with Cyberpunk, running at ultra settings, the desktop system smashed past 60FPS average. It ran 10 FPS faster on average than even the turbo mode from the laptop chip, and 15 FPS faster than standard. That means the desktop card is between 15 and 30% faster, which is actually much closer than I thought it’d be. This sort of difference, at these settings anyway, would be pretty noticeable. Going from a 1% low of just over 30FPS to nearly 50 FPS is a big deal, possibly even more so than the increased average.
Watchdogs is the same story, the desktop card ran just over 10 FPS faster on average than either of the laptop runs making it between 23 and 26% faster. What’s interesting here is the 1% low figure for the desktop card is the same result as the faster laptop run got for it’s average. Again, that would be a noticeable improvement, one you’d actually feel while playing.
Fortnite offers by far the most stark difference. The desktop card offered double the performance of the stock laptop chip, and was 82% faster than the turbo mode result. Hell, it’s 1% low result was almost 20FPS higher than the average of either laptop run. I should make it clear Fortnite is a hard game to truly compare side by side. Exactly what you are doing during the run, how much of the time you spend outside, looking up or down, in gunfights or just running, all drastically affects the in-game performance so take these results with a pinch of salt.
Microsoft flight also shows a decent improvement from the desktop card, again right in the 20-30% fast range. It ran at nearly 36FPS average which doesn’t sound like much, but for Microsoft Flight is (sadly) rather a good result. Of course no matter if you were playing this on the laptop or the desktop you would be turning the settings down a little, but the outright performance gap seems clear.
And lastly in CSGO, that gap is present again. The desktop card offered 22% more performance than the turbo result and 16% more than the stock settings for the laptop – a difference of either 40 or 50 FPS. That is less likely to be noticeable seeing as it’s the difference between 220 and 260 FPS, but it confirms the pattern.
So, the desktop 3060 is between 20 and 30% faster when in a like for like condition, at least compared to this ‘mid tier’ TDP laptop chip. That’s considerably less of a difference than I thought it’d be, although there is one extra benefit the desktop has over the laptop – upgradeability. See, the 10400F I’m using here isn’t exactly the fastest, and the B560 motherboard I’ve got it in supports much faster, even higher core count chips.
Upgrading your CPU, at least at 1080p and in more CPU bound games, can extract more performance out of your graphics card. An example of that is from my 11400F review, where this very 10400F was getting around 84FPS average in Cyberpunk, but the 11600K offers 91FPS average. CSGO was even more stark, the 10400F averaged around 270FPS, the 5600X (which yes isn’t directly compatible as an upgrade, but if you are building from scratch it’s an option) hit nearly 350FPS. The same goes for Watchdogs, with almost 10FPS more from the 11600K than the 10400F. This applies to the 3060 here too, you would get more performance and therefore a more stark gap to the laptop’s chip.
Finally, on the topic of price, assuming you can get the GPU for MSRP which for this Gigabyte OC card I think is somewhere between £400 and £500 – with the standard desktop 3060’s MSRP starting at £300, a desktop build as I have it here with a 1TB M.2 SSD, 10400F and 16GB of 3200MHz RAM would come in at around £1000, or about £300 less than this laptop. Of course that’s without a monitor or peripherals but for £300 I think you’d survive.
Obviously, a laptop and a desktop do serve different purposes and offer different feature sets to each other. The desktop may be upgradable and more performant for the same or less money, but it’s nowhere near as portable or something you can easily use anywhere you are, especially out and about or when travelling. It’s nice to know the laptop chip isn’t hurting too hard from it’s lower power budget, although I still have a rant saved for NVIDIA and their partner’s marketing of these laptop chips that are fundamentally different from their desktop counterparts, yet are labelled EVERYWHERE as the same.
