NVIDIA’s Ada Lovelace GPUs such as the flagship AD102 for GeForce RTX 4090 and RTX L6000 graphics cards are going to be some of the most power-hungry graphics cards ever released. Rumors have already cited insane power draw for the next-gen lineup but Moore’s Law is Dead reiterates that this is indeed the case in his latest video.
Gamers & Workstation Users Get Ready! Next-Gen NVIDIA AD102 GPUs Powering the GeForce RTX 4090 and RTX L6000 To Consume Up To 600W TBP In Reference Designs
From what we know so far, the NVIDIA AD102 GPU is going to be a power-chugging monster & has been reported to feature a TBP of 850W. But as NVIDIA tries to optimize it, the TBP figure is going to come down eventually. As such, the latest TBP figure for NVIDIA’s flagship graphics card is set to 600W (Total Board Power). There are two cards that are going to feature the flagship configuration, the gaming-focused GeForce RTX 4090 and the workstation-focused RTX L6000. These names are not confirmed yet but they are going to be the successors to the RTX 4090 and the RTX A6000 that are based on the GA102 Ampere cores.
NVIDIA GeForce RTX 4090 ‘AD102 GPU’ Gaming Graphics Card With 600W TBP
According to Tom, the NVIDIA GeForce RTX 4090 will have a TBP of 600W at reference and that’s going to go up with the non-reference designs. As we saw with the GeForce RTX 3090 Ti, there are certain models that go beyond the 450W TBP (reference) and up to 516W (custom). The same is going to be the case with Ada Lovelace-based cards. And while this will require lots of cooling, it looks like AIBs are well and ready. We are likely going to see the use of more hybrid-styled GPU coolers rather than traditional air-coolers once we go above the 600W TBP limit. This will mark a 33% increase in TBP over the RTX 3090 Ti.
As for the card itself, it looks like NVIDIA wants to keep the gaming crown itself & the GeForce RTX 4090 could become the first RTX 40 series card to launch as early as Q3 2022. With AMD focusing on getting Navi 33 out first, it looks like that will keep NVIDIA in the gaming throne till Navi 31 comes out & that’s expected to give the green team a serious beating in the efficiency department thanks to its MCM design.
NVIDIA CUDA GPU (RUMORED) Preliminary:
GPU | TU102 | GA102 | AD102 |
---|---|---|---|
Architecture | Turing | Ampere | Ada Lovelace |
Process | TSMC 12nm NFF | Samsung 8nm | TSMC 5nm |
Die Size | 754mm2 | 628mm2 | ~600mm2 |
Graphics Processing Clusters (GPC) | 6 | 7 | 12 |
Texture Processing Clusters (TPC) | 36 | 42 | 72 |
Streaming Multiprocessors (SM) | 72 | 84 | 144 |
CUDA Cores | 4608 | 10752 | 18432 |
L2 Cache | 6 MB | 6 MB | 96 MB |
Theoretical TFLOPs | 16.1 | 37.6 | ~90 TFLOPs? |
Memory Type | GDDR6 | GDDR6X | GDDR6X |
Memory Bus | 384-bit | 384-bit | 384-bit |
Memory Capacity | 11 GB (2080 Ti) | 24 GB (3090) | 24 GB (4090?) |
Flagship SKU | RTX 2080 Ti | RTX 3090 | RTX 4090? |
TGP | 250W | 350W | 450-850W? |
Release | Sep. 2018 | Sept. 20 | 2H 2022 (TBC) |
NVIDIA GeForce RTX L6000 ‘AD102 GPU’ Workstation Graphics Card With 375W TBP
The NVIDIA workstation lineup will also be updated to the new AD102 GPU core and the top part, the RTX L6000 or RTX L8000, will utilize the top AD102 GPU. It is expected to feature a TBP of 320 and up to 375W. According to MLID, current prototypes are being tested with EPS 8-pin connectors though it is very likely that we will see final variants with the new PCIe Gen 5 (16-pin) connectors.
Since the card is a much lower TBP-design compared to the RTX 4090, a dual-slot & blower-style cooler can still sustain it. But this is still a huge 25% increase in power draw for the new graphics card.
Having the flagship GeForce RTX 4090 at 600W TBP means that NVIDIA will also most likely have the GeForce RTX 4080 running at a higher TBP. Previous claims have pointed to a 450W TBP, the same as the current RTX 3090 Ti graphics card.
Plus, graphics cards this power-hungry will make 1000 & 1200W Power Supplies the new standard and that’s also where the new ATX 3.0 platform comes into play. The new ATX 3.0 standard will help deliver sustained power to next-gen PCIe Gen 5 cards that are expected to have ‘Power Excursions’ of up to 1800W. So a 1000-1200W Power Supply will be your best bet if you want to try out the upcoming graphics cards.
NVIDIA Ada Lovelace & Ampere GPU Comparison
Ada Lovelace GPU | SMs | CUDA Cores | Top SKU | Memory Bus | Ampere GPU | SMs | CUDA Cores | Top SKU | Memory Bus | SM Increase (% Over Ampere) |
---|---|---|---|---|---|---|---|---|---|---|
AD102 | 144 | 18432 | RTX 4090? | 384-bit | GA102 | 84 | 10752 | RTX 3090 Ti | 384-bit | +71% |
AD103 | 84 | 10752 | RTX 4070? | 256-bit | GA103S | 60 | 7680 | RTX 3080 Ti | 256-bit | +40% |
AD104 | 60 | 7680 | RTX 4060? | 192-bit | GA104 | 48 | 6144 | RTX 3070 Ti | 256-bit | +25% |
AD106 | 36 | 4608 | RTX 4050 Ti? | 128-bit | GA106 | 30 | 3840 | RTX 3060 | 192-bit | +20% |
AD107 | 24 | 3072 | RTX 4050? | 128-bit | GA107 | 20 | 2560 | RTX 3050 | 128-bit | +20% |