It is known that a smaller
.dll/.so file leads to a faster loading time. For now,
libmxnet.dll is too large and takes too much time to load in my Windows 10.
It seems that if we have a relocatable
.dll/.so file, we could using
nvprune to discard all the useless
CUDA_ARCHes and get a smaller
nvprune, I tried it and it gives me an error:
nvprune fatal : Input file 'libmxnet.dll' not relocatable
libmxnet.so relocatable will provides the possibility to further decrease the size of the
dll file. And may further more decrease the import time.
It seems the symbol are contained in
libmxnet.lib, but I can not merge these symbols into
libmxnet.dll. If someone find a way to redistribute a
libmxnet.dll with symbols, we may take less time import mxnet in python.