Nvidia’s next generation of RTX 30 series graphics cards aren’t out yet, but the company’s data center GPU the A100 has already been benchmarked on Octane Renderer, and it’s officially taken the crown as the world’s fastest GPU with a score of 446 even with RTX off.
The news was shared by Jules Urbach, the CEO of OTOY - the company that developed the Octane Renderer benchmark: “A record breaking week. The @NVIDIA A100 has now become the fastest GPU ever recorded on #OctaneBench: 446 OB4*. #Ampere appears to be ~43% faster than #Turing in #OctaneRender - even w/ #RTX off! (*standard Linux OB4 benchmark, RTX off, recompiled for CUDA11, ref. 980=102 OB)”
It’s unclear exactly what Turing card they are referring to that Ampere seems to be 43% faster, but the RTX 2080 Ti reached a benchmark score of 302 points, making the A100 47.7% faster than the flagship RTX 20 series card.
That’s a pretty incredible achievement for Ampere, and granted it is specifically a data center GPU with a whopping 6912 cuda cores and 54 billion transistors on a 7nm die, along with 40GB of HBME2 memory, so it’s not exactly a consumer-grade card.
However, it does show the increased performance potential of the Ampere architecture, as the A100’s predecessor the Tesla V100 is around 20% slower.
All in all, Nvidia’s next-gen Ampere architecture is shaping up to be something quite interesting, and whilst we don’t have anything official just yet, lots of leaks and rumors are coming out hot and fast for the RTX 30 series, with an expected launch date of September this year. Hopefully we’ll hear something a bit more official from Nvidia themselves soon.
What do you think? Are you excited for Ampere? Which card are you hoping to get? And what would gaming be like with an A100 installed? Let us know your thoughts!