Hello collective wisdom, I'm trying to build FIO with GPU Direct support. My system is a NVIDIA DGX-A100 with CUDA 11.4 and GDS 1.0.2 This is how I'm trying to configure and how it fails. Any ideas what I am doing wrong? export LDFLAGS="-L/usr/local/cuda-11.4/lib64" ./configure --enable-libcufile --enable-cuda --extra-cflags="-I/usr/local/cuda-11.4/include -I/usr/local/cuda-11.4/lib64" config.log shows Compiling test case cuda gcc -D_GNU_SOURCE -include config-host.h -I/usr/local/cuda-11.4/include -I/usr/local/cuda-11.4/lib64 -Werror-implicit-function-declaration -o /tmp/fio-conf--1279005-.exe /tmp/fio-conf--1279005-.c -L/usr/local/cuda-11.4/lib64 -lcuda -lrt -lz -lpthread Compiling test case libcufile gcc -D_GNU_SOURCE -include config-host.h -I/usr/local/cuda-11.4/include -I/usr/local/cuda-11.4/lib64 -Werror-implicit-function-declaration -o /tmp/fio-conf--1279005-.exe /tmp/fio-conf--1279005-.c -L/usr/local/cuda-11.4/lib64 -lcuda -lcudart -lcufile -lcuda -lrt -lz -lpthread /usr/bin/ld: /usr/local/cuda-11.4/lib64/libcufile.so: undefined reference to `dlopen' /usr/bin/ld: /usr/local/cuda-11.4/lib64/libcufile.so: undefined reference to `dlclose' /usr/bin/ld: /usr/local/cuda-11.4/lib64/libcufile.so: undefined reference to `dlerror' /usr/bin/ld: /usr/local/cuda-11.4/lib64/libcufile.so: undefined reference to `dlsym' collect2: error: ld returned 1 exit status -- Thanks, Michael
