Two new next-gen NVIDIA CUDA GPUs have been spotted online and they seem to pack quite a punch. So much so that they obliterate even the NVIDIA RTX 2080 Ti in benchmark scores. The new pair of graphics cards were spotted in the Geekbench database with one of them featuring 118 compute units and the other 108. The two cards are said to have 7,522 and 6,912 CUDA cores respectively.
By the look of things, NVIDIA may be boosting the per-core performance of its next-gen CUDA cards in an attempt to increase efficiency. The Geekbench entries are based on the GPU’s OpenCL benchmark scores with the larger graphics card reaching 184,096 points and the smaller card (not very small) getting 141,654. In comparison, the NVIDIA RTX 2080 Ti maxes out at around 135,000.
|Model||Mystery GPU 1||Mystery GPU 2||Quadro RTX 8000|
|Clock Speed||1110 MHz||1010 MHz||1770 MHz|
|Memory||24 GB||48 GB||48 GB|
As for the clock speed of these so-called NVIDIA next-gen GPUs, they seem to be recorded at 1.11GHz with the smaller one at 1.01GHz. These scores will most likely increase to around 1.5GHz and can be further overclocked by the manufacturer too.
However, the big number here is the memory on each of these cards with the larger 1.11GHz featuring 48GB and the smaller, 1.01GHz packing 24GB. This could mean these cards are not gaming GPUs. With that being said, NVIDIA will most likely use the same architecture from these monster GPUs in order to build their next-gen gaming cards so this is a good sign of things to come.
Keep in mind that these scores could also be completely incorrect but from previous leaks, we doubt it.
Editor for GLITCHED | Geek AF | Gamer | Techie | Foodie | Cars | Not a morning person | Cats Are Better Than Dogs |