backend/cuda: Use NVCC_FLAGS with user-provided NVCC

These flags were ignored when the user specified a compiler other than
the nvcc included in the CUDA installation. Make sure to include them
for consistency. See pmodels/mpich#6954.

Signed-off-by: Ken Raffenetti <raffenet@mcs.anl.gov>

src/backend/cuda/subconfigure.m4

@@ -92,6 +92,7 @@ if test "$with_cuda" != "no" ; then
             # append the flags to be passed to nvcc
             NVCC="$nvcc_bin $NVCC_FLAGS"
         else
+            NVCC="$NVCC $NVCC_FLAGS"
             AC_MSG_WARN([Using user-provided nvcc: '${NVCC}'])
         fi