# NVIDIA GeForce RTX 3050, 18 SMs, sm_86, iters=20000 reps=5
mode       best TF/s   all runs
f16/f16         36.6 36.4/36.6/36.6/36.6/36.6
f16/f32         36.6 36.6/36.6/36.3/36.3/36.3
bf16/f32        36.3 36.3/36.3/36.3/36.3/36.3
s8/s32          72.6 72.6/72.6/72.6/72.6/72.6
dp4a            17.7 17.7/17.7/17.7/17.7/17.7
ffma             8.5 8.5/8.2/8.3/8.3/8.3
