# NVIDIA GeForce RTX 4090, 128 SMs, sm_89, iters=20000 reps=5
mode       best TF/s   all runs
f16/f16        318.9 294.7/302.4/305.9/318.9/318.9
f16/f32        159.5 159.5/156.0/159.5/155.5/155.5
bf16/f32       159.4 159.4/152.7/156.1/152.7/156.0
s8/s32         637.5 637.2/637.5/637.4/637.4/590.3
dp4a           158.5 158.5/134.0/158.5/158.5/158.4
ffma            76.1 65.1/76.0/76.1/76.0/68.9
