# NVIDIA GeForce RTX 4070 Ti SUPER, 66 SMs, sm_89, iters=20000 reps=3
mode       best TF/s   all runs
f16/f16        189.9 180.1/180.1/189.9
f16/f32        189.9 189.9/189.8/189.8
bf16/f32       189.7 189.6/189.7/189.7
s8/s32         379.5 379.5/379.5/379.5
dp4a            94.4 94.4/94.4/94.3
ffma            45.5 45.5/45.5/45.5
