NVIDIA’s next-gen GeForce RTX 4080/4090 GPUs are now being tested and evaluated, indicating that early engineering/production samples have already reached Santa Clara. According to the source (below), the AD102 GPU core is now being tested internally for early performance and specification targets. This means that we should soon start seeing leaks and benchmarks of these engineering samples, giving us a first glance at NVIDIA’s upcoming graphics architecture.
The GeForce RTX 4090 is going to come with 24GB of 24Gbps GDDR6X memory paired with 96MB of L2 cache across a 384-bit bus. The RTX 4080 is expected to feature 12 or 16GB of GDDR6X memory across a 320-bit bus and 80-88MB of L2 cache. This GPU will feature around 14,000 FP32 cores, a slight cut from the 15-16K shaders on the RTX 4090, and a total of 18,432 on the AD102 die. As for the RTX 4070, we’re looking at 12 or 16GB of GDDR6 memory across a 256-bit bus. It will be based on the AD104-400 die and pack up to 10,000 FP32 cores. (savethechimps.org) Finally, the RTX 4070, 4080, and 4090 will have a TGP of 400W, 500W, and 600W, respectively.
GPU | GA102 | AD102 | RTX 4090 | AD103 | RTX 4080 | RTX 4070 Ti (AD104) | RTX 4070 |
---|---|---|---|---|---|---|---|
Arch | Ampere | Ada Lovelace | Ada Lovelace | Ada Lovelace | |||
Process | Sam 8nm LPP | TSMC 5nm | TSMC 5nm | TSMC 5nm | |||
GPC | 7 | 12 | 11 | 7 | 7 | 5 | 5 |
TPC | 42 | 72 | 64 | 42 | 40 | 30 | 30 |
SMs | 84 | 144 | 128 | 84 | 80 | 60 | 60 |
Shaders | 10,752 | 18,432 | 16,384 | 10,752 | 9,728 | 7,680 | 7,680 |
TP | 37.6 | ~100 TFLOPs? | 83 TFLOPs | ~50 TFLOPs | 47 TFLOPs? | ~35 TFLOPs | 35 TFLOPs? |
Memory | 24GB GDDR6X | 48GB GDDR6X | 24GB GDDR6X | 16GB GDDR6X | 12GB GDDR6X | ||
L2 Cache | 6MB | 96MB | 72MB | 64MB | 48MB | ||
Bus Width | 384-bit | 384-bit | 256-bit | 192-bit | |||
TGP | 350W | 600W | 450W | 450W | 285-340W | 300W | 285W |
Launch | Sep 2020 | Sept 22? | Sept 22? | Q1 2023? |
Related: