The power limits of NVIDIA's Ada Lovelace GPUs that will be powering the next-generation GeForce RTX 40 graphics cards have been revealed by Kopite7kimi. The power limits include both desktop & mobile parts.
NVIDIA GeForce RTX 40 Desktop & Mobile GPU Power Limits Detailed: Ada Lovelace Powered AD102 Flagship Desktop Maxes Out at 800W, Flagship Mobile AD103 Maxes Out at 175W
The NVIDIA GeForce RTX 40 series lineup based on the Ada Lovelace graphics architecture will come in both desktop and mobile flavors. We have heard various reports of how the flagship is going to feature a TGP of up to 900W but as we know, the power limits are the max the GPU can reach and are not reflective of the actual TGP/TBP figures. These limits may only be reachable by heavily overclocked and custom-cooled variants.
A truth. The power limits:
AD102, 800W;
AD103 (DT), 450W, AD103 (Mobile), 175W;
AD104 (DT), 400W, AD104 (Mobile), 175W;
AD106 (DT), 260W, AD106 (Mobile), 140W.
But I don't think we need to use the full power cap.— kopite7kimi (@kopite7kimi) June 18, 2022
So coming to the power limits, we have four NVIDIA Ada Lovelace GPU SKUs, the AD102, AD103, AD104, and AD106. All GPUs except the AD102 will receive a mobile variation and the reason is simply that the flagship chip has much higher power requirements that the laptop platform can't support. The same was the case with the GA102 GPU so the GA103 became the fastest mobile offering.
Starting with the desktop lineup, the NVIDIA AD102 GPU will feature an 800W power limit, the AD103 GPU will feature a 450W power limit, the AD104 GPU will feature a 400W power limit & AD106 will feature a 260W power limit. As for mobile parts, the AD103 mobile GPU will feature a 175W power limit, AD104 mobile GPU will feature a 175W power limit and the AD106 mobile will rock a 140W power limit. Rounding up the figures, we have:
NVIDIA Ada Lovelace Desktop GPUs:
- AD102 - 800W Max Power Limit
- AD103 - 450W Max Power Limit
- AD104 - 400W Max Power Limit
- AD106 - 260W Max Power Limit
NVIDIA Ada Lovelace Mobile GPUs:
- AD103 - 175W Max Power Limit
- AD104 - 175W Max Power Limit
- AD106 - 140W Max Power Limit
This is a drastic increase in the power limits for the desktop and mobile parts. The AD102's 800W power limit is already bonkers but the actual graphics card that utilizes it will end up with a 600W TGP. The AD103 and AD104 GPUs are very similar to each other so the TGP figures are almost the same on mobile and a 50W difference on the desktop segment. The AD106 is in the mainstream 260W segment while the actual graphics cards using it will end up close to 200W TGPs. The mobile parts max out at 175W power limits because that's the best the platform could do in the given area. Anything more would require a major redesign of existing laptops and coolers for them and it looks like most manufacturers don't want to make that extra investment right now.
Based on the power limit differences between the desktop and mobile parts, there's once again going to be a huge performance difference between the two NVIDIA GeForce RTX GPU segments. But that's only due to the limitations on the mobile platform, still, we expect to see a major boost when it comes to performance on the mobile side while desktops take things to the next level.
NVIDIA Ada Lovelace 'GeForce RTX 40' GPU Configurations
| GPU Name | GPCs / TPCs | SMs Per TPC / Total | CUDA Cores | L2 Cache | Memory Bus | Power Limit Desktop (Peak) | Power Limit Mobile (Expected) |
|---|---|---|---|---|---|---|---|
| AD102 | 12 / 6 | 2 / 144 | 18432 | 96 MB | 384-bit | 800W | N/A |
| AD103 | 7 / 6 | 2 / 84 | 10752 | 64 MB | 256-bit | 450W | 175W |
| AD104 | 5 / 6 | 2 / 60 | 7680 | 48 MB | 192-bit | 400W | 175W |
| AD106 | 3 / 6 | 2 / 36 | 4608 | 32 MB | 128-bit | 260W | 140W |
| AD107 | 3 / 4 | 2 / 24 | 3072 | 32 MB | 128-bit | 180W | 80W |
