Later this year NVIDIA’s Ada Lovelace GPUs will be available to gamers in the next-gen GeForce RTX graphics cards. These GPUs are anticipated to pack with some extreme specs and some extreme GPU requirements to operate as described by Kopite7kimi in his latest news of leaks.
There might be a new rumor covering the naming pattern used by the next-gen NVIDIA GeForce graphics cards with some now indicating a change in name to GeForce RTX 50 series instead of the GeForce RTX 40 series which was expected earlier. In the end, whatever NVIDIA chooses to call the next GeForce lineup isn’t the main highlight here but its underlying specifications are.
NVIDIA GeForce RTX 4090 (or RTX 5090) ‘Ada Lovelace’ Graphics Card With GA102-300 GPU ‘Rumored’ Specs
We know from the leaked NVIDIA Ada Lovelace GPU configurations which indicate that the flagship GA102 Gaming chip will be packed with a total of 144 SMs, a 71% increase over the current GA102 GPU, and comes with a huge 18,432 CUDA core count. The chip used by the flagship GeForce graphic cards will come with the AD102-300 GPU and we can’t tell if this will be the full-fat SKU since the RTX 3090 Ti comes with the GA102-350 configuration and the NVIDIA GeForce RTX 3090 which has GA102-300 SKU is a cut-down SKU.
According to Kopite7kimi, the flagship NVIDIA GeForce graphics card will support PG137/139-SKU30 PCB. These are both reference and Founders Edition designs. The reference design is provided to AIB partners as a reference for their custom models while the FE design is exclusive to the reference NVIDIA’s Founders Edition model.
The graphics card is said to pack with 24 GB memory capacity with a memory speed of 21 Gbps across a 384-bit bus interface. The memory design is the same as the flagship Ampere GeForce, the GeForce RTX 3090 Ti, & should come with the same amount of bandwidth at 1004 Gb/s. A single PCIe Gen 5 connector with a TGP rating of up to 600W will be the power input. You will need a 4x 8-pin to 1x 16-pin adapter if you are using an older PSU or a single Gen 5 to Gen 5 connector if you have a supporting power supply.
NVIDIA Ada Lovelace ‘GeForce RTX 40’ (or GeForce RTX 50) GPU Configurations
GPU NAME | GPCS / TPCS | SMS PER TPC / TOTAL | CUDA CORES | 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 between NVIDIA Ada Lovelace & Ampere GPU
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% |
NVIDIA Ada Lovelace GPU family is expected to bring a huge performance jump similar to Maxwell to Pascal. It is expected to launch in the second half of the year 2022 but supply and pricing are expected to be similar to existing cards despite NVIDIA spending billions of dollars to acquire those good TSMC 5nm wafers.
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) |