We’ve been hearing rumors of a Super variant of the Turing flagship for a while now, but nothing concrete has surfaced as of now. Today, a well-known source has once again rattled the cage, this time with the specifications of the GeForce RTX 2080 Ti Super. This is the same source that first leaked info about the RTX 20 Super and the GTX 16 series lineup, so we’re forced to pay attention.
NVIDIA GeForce RTX 2080 Ti Super: 4608 Cores & 16GB GDDR6 Memory
According to the new info, the RTX 2080 Ti Super will feature the full TU102 die, packing as many as 4608 cores and 16GB of GDDR6 memory potentially via a 384-bit bus. This would make it considerably faster than the vanilla thanks to a healthy bandwidth boost as well as better shader capabilities.
| Graphics Card Name | NVIDIA GeForce RTX 2080 Ti SUPER | NVIDIA GeForce RTX 2080 Ti |
|---|---|---|
| GPU | TU102 | TU102 |
| Process | 12nm | 12nm |
| CUDA Cores | 4608 | 4352 |
| Memory | Up To 16 GB GDDR6 | Up To 11 GB GDDR6 |
| Memory Speed | 16 Gbps | 14.00 Gbps |
| Memory Interface | 384-bit | 352-bit |
| Power Connectors | 8+8 Pin | 8+8 Pin |
| TDP | 300W? | 260W (Founders) 250W (Reference) |
| Price | $999 | $999 (Launch) |
Comparing the specifications, we should see an overall performance boost of the same magnitude as the lower-end Super cards, somewhere in the 15-20% range. However, at present all this is a rumor and might not pan out as expected.
NVIDIA is rumored to launch the Ampere lineup based on the 7nm node by the end of 1H 2020, and there’d be no point to release a 2080 Ti Super if it’s only marginally faster than the RTX 3080, not to mention it’s also a mouthful. We’ll keep you posted.