hiperc
hiperc copied to clipboard

Published 20 hours ago •

Reame
Issues

unroll convolution

Open tkphd opened this issue 6 years ago • 0 comments

Inner loops on CUDA convolution code should run faster using a #pragma unroll statement.

#pragma unroll N in CUDA
#pragma unroll N in OpenCL
Unavailable for OpenAcc, but try -Munroll flag with pgcc
#pragma unroll N for icc

Dec 20 '17 20:12 tkphd