I am working on a research project and we are trying to optimize the network communication in a distributed PyTorch deployment. I modified the gloo library and I am able to compile and run programs that use this library (similar to what you can find in gloo/gloo/examples
).
So I replaced the third_party/gloo
folder in PyTorch with my version of gloo and I am trying to compile it.
The tricky part is that my version of gloo requires some additional libraries and special linker options, and I cannot find the right place where to specify these options.
I tried to add these options to the Caffe2_DEPENDENCY_LIBS
variable in pytorch/caffe2/CMakeLists.txt
with no success.
My compilation stops with the linker error:
/pytorch/build/lib/libcaffe2_gpu.so: undefined reference to <my code>
/pytorch/build/lib/libcaffe2.so: undefined reference to <my code>
Can someone point me in the right direction?
Thanks!