As NVIDIA’s RTX 30 series “Ampere” launch draws near, we’ve started hearing more and more about the next-gen GeForce lineup. Today’s leak comes as a confirmation of sorts. Reputed Twitter source, _rogame has reported that NVIDIA is presently testing three GPUs based on the GA102 die with 10GB, 12GB, and 24GB of memory. These will likely form the top-tier Ampere consumer cards, namely the RTX 3080, 3080 Ti, and the new Titan.
| Part | PCB | Chip | Model | Extension | Memory | Interface | TBP | Connectors |
| SKU10 | PG132 | GA102 | RTX 3090 | (Ti/Super)* | 24 GB GDDR6X (Double-Sided) | 384-bt | 350 W | 3x DP, HDMI NVLink |
| SKU20 | PG132 | GA102 | RTX 3080 | (Ti/Super)* | 11/12 GB GDDR6X* | 352-bit* | 320 W | 3x DP, HDMI |
| SKU30 | PG132 | GA102 | RTX 3080 | none | 10 GB GDDR6X | 320-bit | 320 W | 3x DP, HDMI |
This is in line with what Igor shared a while back. As per his sources, the RTX 3080 will feature a 320-bit bus, the 3080 Ti a 352-bit bus, and the RTX Titan|3090 will be paired with a 384-bit bus. All three GPUs will pack the lastest GDDR6 (or GDDR6X) memory standard with a TDP of 325W for the 3080 series and 350W for the Titan.
This report from _rogame confirms what Igor reported a month back. The top three Ampere GPUs will be based on the same GA102 die with the flagship Titan GPU packing more than 5,000+ CUDA cores. The RTX 3080 will likely feature around 4,350 shaders (as indicated in recent rumors) and the 3080 Ti will slot in between the two with slightly fewer cores than the Titan and a trimmed bus.
I’m curious how the pricing of NVIDIA’s next-gen Ampere graphics cards will be like. Considering that the RTX 2080 Ti, despite being just marginally faster than the RTX 2080 Super costs north of $1,000, I really hope Jensen goes easy on us. Either way, with Navi 2x and Big Navi primed for a launch in the same quarter (Q4 2020), the prices will come down sooner or later.
