At its GPU Technology Conference (GTC) 2022 event, Nvidia released its RTX 40-series desktop GPUs based on the Ada Lovelace architecture. The lineup consisted of the $1,599 RTX 4090 flagship, alongside two RTX 4080 SKUs, the RTX 4080 16GB model for $1,199 and the 4080 12GB model for $899.
However, the VRAM capacity and bandwidth isn't the only thing different between the two cards as Nvidia has cut down the CUDA core count of the 12gig pretty significantly compared to the 16GB 4080 model. Overall, it is a reduction of more than 20% for the core count. Meanwhile, the bandwidth takes an even bigger hit as it is slashed down by more than 30%.
While Nvidia has done this type of naming before, for example with the GTX 1050 3GB model, or with GTX 1060 3GB variant, the core count difference, and the overall performance difference between two such variants has been never been this big. Hence, the company has rightly received criticism this time over this 4080 12GB situation from media and enthusiasts since naming such a cut-down product the same as the 4080 16GB model makes little to no sense, and is extremely misleading.
Nvidia seems to have realized the issue and has pulled the launch of the RTX 4080 12GB model, which was supposed to launch next month on the 16th. The company writes:
The RTX 4080 12GB is a fantastic graphics card, but it’s not named right. Having two GPUs with the 4080 designation is confusing.
So, we’re pressing the “unlaunch” button on the 4080 12GB. The RTX 4080 16GB is amazing and on track to delight gamers everywhere on November 16th.
For now, it is unclear what Nvidia will call the 4080 12GB. Perhaps it will settle with something like the RTX 4080 Lite edition or maybe an RTX 4070 Ti.