# NVIDIA GeForce RTX 3050, 18 SMs, sm_86, iters=20000 reps=5
mode       best TF/s   all runs
f16/f16         36.6 36.6/36.6/36.6/36.6/36.6
f16/f32         18.3 18.3/18.3/18.2/18.2/18.2
bf16/f32        18.2 18.2/18.2/18.2/18.2/18.2
s8/s32          72.8 72.8/72.8/72.6/72.3/72.3
dp4a            17.7 17.7/17.7/17.7/17.7/17.7
ffma             8.4 8.4/8.3/8.3/8.3/8.3
