Because they’re running it at Ultra on the lowest end hardware configuration, with an x86_64 to arm64 translation, and a shader translation and a windows api wr

Because they’re running it at Ultra on the lowest end hardware configuration, with an x86_64 to arm64 translation, and a shader translation and a windows api wrapper. For comparison, the M1 was routinely benchmarked in NVIDIA 1050 TI to 1070 territory depending on the benchmark. This is only a few fps behind native.