If you are planning to buy a gaming laptop in the coming months based on the new NVIDIA GPUs, based on its GeForce Ampere architecture, and under the trade name RTX 3000 Mobile, then you have to know that not all variants of the RTX 3000 for laptops are equal. That is why we have ordered the information so that you know the GPU that your future laptop carries.
In the past, CES NVIDIA presented its new RTX 3000 Mobile, based on its latest generation architecture, and a few weeks later we already have the first laptops with them, however not all RTX 3080 Mobile, RTX 3070 Mobile, and RTX 3060 Mobile do not They are equal.
NVIDIA's new generation of GPUs for laptops bring with them all the benefits of their full desktop counterparts, plus third-generation Max-Q technology.
The variants of the NVIDIA GeForce RTX 3000 Mobile
The two observations that can be made after a review of all the information on the variants of the RTX 3000 Mobile are two, the first is that each variant has a TGP 5 W lower than the previous variant, with the odd exception, and the second that the Max-Q models use 12 Gbps memory.
To facilitate the information we have divided the different models by commercial name, that is, a table for RTX 3080 Mobile, a second for RTX 3070 Mobile, and a third and last table for the RTX 3060 Mobile.
So we can find that two laptops in theory with the same GPU do not have the same variant and therefore are different in clock speed and therefore in power and performance.
Variants of the NVIDIA RTX 3080 Mobile
Variant | CUDA cores | Base MHz | MHz Boost | TFLOPS | VRAM bus | Gbps | TGP (W) |
RTX 3080 Mobile | 6144 | 1350 | 1710 | 16.6 | 256 bits | 14 | 150 |
RTX 3080 Mobile | 6144 | 1320 | 1695 | 16.2 | 256 bits | 14 | 145 |
RTX 3080 Mobile | 6144 | 1275 | 1665 | 15.7 | 256 bits | 14 | 140 |
RTX 3080 Mobile | 6144 | 1260 | 1665 | 15.5 | 256 bits | 14 | 135 |
RTX 3080 Mobile | 6144 | 1230 | 1635 | 15.1 | 256 bits | 14 | 130 |
RTX 3080 Mobile | 6144 | 1185 | 1605 | 14.6 | 256 bits | 14 | 125 |
RTX 3080 Mobile | 6144 | 1155 | 1575 | 14.2 | 256 bits | 14 | 120 |
RTX 3080 Mobile | 6144 | 1110 | 1545 | 13.6 | 256 bits | 14 | 115 |
RTX 3080 Mobile Max-Q | 6144 | 930 | 1365 | 11.4 | 256 bits | 14 | 90 |
RTX 3080 Mobile Max-Q | 6144 | 870 | 1320 | 10.7 | 256 bits | 12 | 85 |
Variants of the NVIDIA RTX 3070 Mobile
Variant | CUDA cores | Base MHz | MHz Boost | TFLOPS | VRAM bus | Gbps | TGP (W) |
RTX 3070 Mobile | 5120 | 1215 | 1620 | 12.4 | 256 bits | 14 | 125 |
RTX 3070 Mobile | 5120 | 1170 | 1590 | 12 | 256 bits | 14 | 120 |
RTX 3070 Mobile | 5120 | 1110 | 1560 | 11.4 | 256 bits | 14 | 115 |
RTX 3070 Mobile Max-Q | 5120 | 930 | 1410 | 9.5 | 256 bits | 12 | 90 |
RTX 3070 Mobile Max-Q | 5120 | 855 | 1365 | 8.8 | 256 bits | 12 | 85 |
Variants of the NVIDIA RTX 3060 Mobile
Variant | CUDA cores | Base MHz | MHz Boost | TFLOPS | VRAM bus | Gbps | TGP (W) |
RTX 3060 Mobile | 3840 | 1387 | 1702 | 10.7 | 192 bits | 14 | 115 |
RTX 3060 Mobile | 3840 | 1342 | 1680 | 10.3 | 192 bits | 14 | 110 |
RTX 3060 Mobile | 3840 | 1305 | 1642 | 10 | 192 bits | 14 | 105 |
RTX 3060 Mobile | 3840 | 1267 | 1605 | 9.7 | 192 bits | 14 | 100 |
RTX 3060 Mobile | 3840 | 1215 | 1567 | 9.3 | 192 bits | 14 | 95 |
RTX 3060 Mobile | 3840 | 1163 | 1530 | 8.9 | 192 bits | 14 | 90 |
RTX 3060 Mobile | 3840 | 1035 | 1485 | 7.9 | 192 bits | 14 | 85 |
RTX 3060 Mobile | 3840 | 900 | 1425 | 6.9 | 192 bits | 14 | 80 |
RTX 3060 Mobile Max-Q | 3840 | 1050 | 1402 | 8.1 | 192 bits | 12 | 70 |
RTX 3060 Mobile Max-Q | 3840 | 975 | 1357 | 7.5 | 192 bits | 12 | 65 |
RTX 3060 Mobile Max-Q | 3840 | 817 | 1282 | 6.3 | 192 bits | 12 | 60 |
Why are there different variants?
We have explained the reason for this many times, not all the chips on a wafer come out the same, some are one hundred percent functional while others are only partially, this causes different variations of a chip that correspond to different models from the same range.
But then there are what are known as parametric yields, these are discovered in the testing phase and they are fully functional chips within specifications but at the speed and consumption ratio, they are not all symmetrical. Which is not a problem for desktop graphics cards.
In the case of GPUs for laptops in which the thermal and consumption specifications influence much more, it is where these variations make much more sense, in order to ensure optimal performance of the laptop, in which a simple variation of 5 W in the TGP of the graphic subsystem can lead to problems from a heat and consumption perspective.
0 Comments