We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
#cuda optimize examples
reachs 78% of the FP32 peak performance on Tesla T4 GPU.