NVIDIA’s Ada Lovelace GPUs will be coming to gamers later this year in next-gen GeForce RTX graphics cards. These cards would have to rock some extreme specs and some extreme GPU requirements to perform as detailed by Kopite7kimi in his latest series of leaks.
NVIDIA’s next-gen flagship GeForce graphics card detailed: AD102 GPU, 24GB memory and 21Gbps, TGP up to 600W
There could be a new rumor surrounding the naming convention used by next-gen GeForce graphics cards, with some now pointing to a renaming to the GeForce RTX 50 series instead of the previously expected GeForce RTX 40 series. In the end, not everything NVIDIA decides to call the next GeForce line is as important as its underlying specs.
**90 PG137/139-SKU330, AD102-300, 24G 21Gbps, 600W
– kopite7kimi (@kopite7kimi) April 11, 2022
NVIDIA GeForce RTX 4090 ‘Ada Lovelace’ Graphics Card with Rumored Specs GA102-300 GPU
We know from leaked NVIDIA Ada Lovelace GPU configurations which show that the flagship GA102 Gaming chip will rock a total of 144 SMs, a 71% increase over the existing GA102 GPU, and house a massive 18,432 cores. CUDA. The chip used by the flagship GeForce graphics card will rock the AD102-300 GPU and we can’t say for sure if it will be the full SKU since the RTX 3090 Ti with the GA102-350 rocks the full config while the GA102 -300 SKU featured on the NVIDIA GeForce RTX 3090 is a discounted SKU.
According to Kopite7kimi, the GeForce flagship graphics card will be based on the PG137/139-SKU30 PCB. These are both reference models and Founders Edition. The reference design is provided to AIBs as a reference for their custom models, while the FE design is exclusive to the Founders Edition reference model.
The graphics card would retain the 24 GB memory capacity while offering 21 Gbps memory speed over a 384-bit bus interface. It’s the same memory design as the Ampere GeForce flagship, the GeForce RTX 3090 Ti, and should bump the same amount of bandwidth at 1004 Gbps. As for the power input, the card is said to feature a single PCIe Gen 5 slot with a TGP power of up to 600W. You will need a 4x 8 pin to 1x 16 pin adapter if you are using a block older power supply or a single Gen 5 to Gen 5 connector if you have a compliant power supply.
NVIDIA Ada Lovelace ‘GeForce RTX 40’ GPU Configurations
| GPU name | GPC / TPC | SM by TPC / Total | CUDA Colors | L2 Cache | Memory bus |
|---|---|---|---|---|---|
| AD102 | 12 / 6 | 2 / 144 | 18432 | 96 MB | 384 bit |
| AD103 | 7 / 6 | 2 / 84 | 10752 | 64 MB | 256 bit |
| AD104 | 5 / 6 | 2 / 60 | 7680 | 48 MB | 192 bit |
| AD106 | 3 / 6 | 2 / 36 | 4608 | 32 MB | 128 bit |
| AD107 | 3 / 4 | 2 / 24 | 3072 | 32 MB | 128 bit |
Comparison of NVIDIA Ada Lovelace and Ampere GPUs
| There is a Lovelace GPU | SMS | CUDA Colors | Top SKU | Memory bus | GPU amp | SMS | CUDA Colors | Top SKU | Memory bus | SM increase (% over amp) |
|---|---|---|---|---|---|---|---|---|---|---|
| AD102 | 144 | 18432 | RTX4090? | 384 bit | GA102 | 84 | 10752 | RTX-3090Ti | 384 bit | +71% |
| AD103 | 84 | 10752 | RTX 4070? | 256 bit | GA103S | 60 | 7680 | RTX-3080Ti | 256 bit | +40% |
| AD104 | 60 | 7680 | RTX 4060? | 192 bit | GA104 | 48 | 6144 | RTX-3070Ti | 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% |
The NVIDIA Ada Lovelace family of GPUs is expected to bring a similar generational leap from Maxwell to Pascal. It’s expected to launch in the second half of 2022, but expect supply and pricing to be similar to current cards despite NVIDIA spending billions of dollars to get those good good 5nm TSMC wafers .
NVIDIA CUDA GPU (RUMOR) Preview:
| GPUs | 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 Colors | 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-2080Ti | RTX-3090 | RTX4090? |
| TGP | 250W | 350W | 450-850W? |
| Release | September 2018 | September 20 | 2H 2022 (to be confirmed) |