The Ashes Escalation benchmark of NVIDIA’s Ampere flagship, the GeForce RTX 3080 has surfaced, giving us an idea of the performance of the RTX 30 series GPUs in real-world gaming workloads. In this benchmark, the RTX 3080 was paired with an Intel Core i9-9900K and tested at the 4K Crazy preset using the DX12 API.

The result was an average of 88.3 FPS, and a score of 8,700 points. Luckily, we recently tested the Turing heavy-weights using the same settings and can now compare them to the RTX 3080 in this benchmark:

As you can see, the GeForce RTX 3080 is roughly 25% faster than the RTX 2080 Ti and approximately 50% faster than the RTX 2080 Super in Ashes Escalation at 4K. It’s interesting to note that AoTS features heavy use of async compute, and as such, this should be viewed as one of the worst performing tests for Ampere. This is because the GA102 removes the dedicated INT32 pipeline and replaces it with a mixed FP32/INT32 datapath which primary handles compute workloads.
Regardless, this means that the RTX 3080 should be around 25% faster than the RTX 2080 Ti and 50% faster than the RTX 2080 Super at the very least. Not very impressive figures, but this is just one benchmark.