The power settings for NVIDIA GeForce RTX 30 mobility GPUs that include the GeForce RTX 3080, RTX 3070 and RTX 3060 have been detailed. The NVIDIA GeForce RTX 30 series will be launched in the Max-Q and Max-P versions, featuring different core configurations, clocks and memory.
NVIDIA GeForce RTX 30 Mobility GPU Leaky power configurations, GeForce RTX 3080 up to 165 W for notebook enthusiasts
There are three power profiles on which NVIDIA GeForce RTX 30 Mobility GPUs will operate. These include the base TGP (Total Graphics Power), cTGP (Configurable Total Graphics Power) and PPAB, which is an AI-accelerated dynamic boost range. PPAB will be optimized primarily by OEM laptop manufacturers for their solutions based on cooling capabilities.
For Max-P variants, PPAB would provide 15 Watts of additional power for the GPU, while in Max-Q variants, this would be rated at up to 20 Watts. The reason that Max-Q variants offer a higher PPAB range is that they have lower TGP in the first place, while Max-P variants come with higher minimum TGPs.
There is now a report by NotebookCheck that NVIDIA has removed the Max-Q and Max-P tags from their GPUs and that OEMs are free to decide which TGPs they want to define for their notebooks running GeForce RTX 30 Mobility GPUs. This means that the minimum TGP must be intact, but other metrics, such as cTGP and PPAB ranges, will be entirely dependent on what OEMs plan to do with them.
NVIDIA GeForce RTX 3080 16 GB Mobility GPU – Laptops from $ 1,999
The NVIDIA GeForce RTX 3080 graphics card will be the flagship of the NVIDIA RTX 30 series notebook GPU family. The NVIDIA GeForce RTX 3080 Mobility will feature 48 SMs in the GA104 GPU SKU (GA104-775). This chip would have a total of 6144 cores, a base clock of 1245 MHz and a boost clock of 1710 MHz.
The video card also has 16 GB VRAM that works on a 256-bit bus interface, but it is not yet known whether NVIDIA is using 12 or 14 Gbps DRAM modules for its mobility line. In the case of 14 Gbps, the GPU will be able to offer a maximum bandwidth of 448 GB / s.
The Max-Q variant of the GeForce RTX 3080 Mobility GPU will have an initial 80W TGP, which will be configurable up to 165W. GeForce RTX 3080 laptops will start at $ 1,999 and offer 100+ FPS at 1440p resolution across multiple game titles.
NVIDIA GeForce RTX 3070 8 GB Mobility GPU – Laptops starting at $ 1299 in the USA
The NVIDIA GeForce RTX 3070 Mobility will feature 40 SMs in the GA104 GPU SKU (presumably the GA104-770). This chip would have a total of 5120 cores, a base clock of 1290 MHz and a boost of up to 1620 MHz. The video card also has 8 GB VRAM that works on a 256-bit bus interface, but it remains to be seen whether NVIDIA is using 12 or 14 Gbps DRAM modules for its mobility line. In the case of 14 Gbps, the GPU will be able to offer a maximum bandwidth of 448 GB / s.
The TGPs for the GeForce RTX 3070 will feature a basic 80W TGP as well, but can only be configured up to 150W. GeForce RTX 3070 laptops will start at $ 1299 and offer 90 FPS 1440p for gaming.
NVIDIA GeForce RTX 3060 6 GB Mobility GPU – Laptops from $ 999
Finally, we have the 6 GB GeForce RTX 3060 which was also announced for the desktop segment today, but with 12 GB of GDDR6 memory. Both the desktop and mobility variants feature the GA106 GPU, but the notebook variant comes with 30 SM units enabled, resulting in a total of 3840 CUDA cores. The graphics chip will feature clocks starting at 1283 MHz and going up to 1703 MHz boost frequencies.
As mentioned before, the 6 GB GeForce RTX 3060 graphics card will come with a total of 6 GB of GDDR6 memory running on a 192-bit bus interface. The speeds of the memory pins must be set to 12 Gbps, which should result in a total bandwidth of 288 GB / s. The 6 GB GeForce RTX 3060 will have 60W and up to 130W TGPs.
According to NVIDIA itself, RTX 3060 laptops will start at $ 999 and offer 30% better performance than Sony’s PS5 console. It appears that NVIDIA is targeting the laptop as a 1080p 90 FPS gaming solution.
NVIDIA GeForce RTX 30 Mobility GPU Line:
| GPU Name | NVIDIA GeForce RTX 3060 | NVIDIA GeForce RTX 3070 | NVIDIA GeForce RTX 3080 |
|---|---|---|---|
| Process Node | Samsung 8nm | Samsung 8nm | Samsung 8nm |
| SKU GPU | GA106 | GA104-770 | GA104-775 |
| SMS | 30 | 40 | 48 |
| CUDA Cores | 3840 | 5120 | 6144 |
| Base Clock | 1283 MHz | 1290 MHz | 1245 MHz |
| Boost Clock | 1703 MHz | 1620 MHz | 1710 MHz |
| Memory clock | 12 Gbps | 12 Gbps | 12 Gbps |
| Memory Type | GDDR6 | GDDR6 | GDDR6 |
| Memory size | 6 GB | 8 GB | 8/16 GB |
| Memory Bus | 192 bits | 256 bits | 256 bits |
| Bandwidth | 288 GB / s | 384 GB / s | 384 GB / s |
| TGP | 60-115W | 80-125W | 80-150W + |
| settings | Max-Q Max-P |
Max-Q Max-P |
Max-Q Max-P |
| Launch | Q1 2021 | Q1 2021 | Q1 2021 |
News source: HXL