How Nvidia RTX 30-series GPUs make laptop shopping much harder


Nvidia’s next-gen graphics have finally arrived in notebooks, and as you can see in our GeForce RTX 3080 laptop GPU review, the mobile version of Team Green “Ampere” architecture spits out frames at blazingly fast rates. But the performance you get from mobile RTX 30-series graphics chips won’t necessarily be equal, even between two laptops with the same hardware inside.

One GeForce RTX 3070 laptop might not be as fast as another, for example. And it’s even possible that in some cases a slower GPU could theoretically be faster than pricier models. Yes, you might (might) see a GeForce RTX 3070 laptop that can match or possibly outperform an RTX 3080 laptop in some games, though we’ll need to see more reviews trickle in before we can state that definitively. Regardless, you’re going to want to do a lot more homework than before when you’re shopping for a gaming laptop.

It’s all because of some fundamental changes Nvidia made with the specifications for the GeForce RTX 30-series mobile GPUs. Here’s what you need to know.

GeForce RTX 30-series desktop vs. laptop cores

Nvidia

The laptop versions of Nvidia’s Ampere-based RTX 30-series GPUs don’t match the specs of the desktop models.

Before we dig deeper into wattage and clock speeds—the real thing to look for in Nvidia gaming laptops now—it’s worth noting that these GeForce RTX 30-series GPUs aren’t the same ones found in desktop graphics cards. The laptop versions of the RTX 3060, 3070, and 3080 don’t have the same number of CUDA graphics cores as the desktop models.

  • RTX 3080: 8,704 desktop CUDA cores versus 6,144 mobile CUDA cores
  • RTX 3070: 5,888 desktop CUDA cores versus 5,120 mobile CUDA cores
  • RTX 3060: 3,584 desktop CUDA cores versus 3,840 mobile CUDA cores

That’s a big change from the last two generations of Nvidia’s mobile chips. In the GTX 10-series and RTX 20-series, the laptop versions of the GPUs were identical to their desktop siblings, but simply tuned to draw less power and hit lower clock speeds. No more.

Nvidia didn’t tweak the mobile RTX 30-series GPUs in an ordered way, either. The mobile GeForce RTX 3080 takes a massive hit compared to the desktop version, shedding well over 2,500 CUDA cores. On paper, the RTX 3080 found in gaming laptops has a lot more in common with the RTX 3070 desktop card. The mobile RTX 3070 took a sizeable but not-quite-as-ludicrous hit to its own CUDA core count, while the notebook version of the RTX 3060 actually has more cores than the desktop graphics card. Madness.

These are still great mobile graphics options, don’t get me wrong. Just don’t expect the same performance out of a five-pound gaming laptop with a cut-down version of the RTX 3080 that you’d get from a hulking triple-slot RTX 3080 desktop card. If you want to know more, our GeForce RTX 3080 laptop GPU review dives deeper into the differences.

Wattage and the muddying of Max-Q

rtx 30 laptops Nvidia

The bigger issue for buyers is the extreme flexibility that laptop makers have when configuring an RTX 30-series GPU inside their offerings. As you can see in the chart above, the minimum and maximum GPU subsystem power limits vary wildly for each GPU, as do the available clock speed options. Those are interrelated, as the amount of power pumped into a GPU heavily affects clock speeds and overall performance. But laptops need cooling systems capable of dissipating the heat that comes from that, which is the reason Nvidia allows the wide range of configurations available per GPU. Rating the GeForce RTX 3080 mobile chips for 80 watts to 150+ watts lets the chip slip into thin-and-light laptops and hulking desktop replacements alike.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: