-MP compiler flag returns unknown option error

Accelerated Computing CUDA CUDA on Windows Subsystem for Linux
1

In my Makefile I want to compile cuda source files with dependency generation. The rule

%.o : %.cu
        $(NVCC) $(NVCCFLAGS) -c $< -o $@ -MMD

works just fine, but notice I don’t include -MP flag (NVCC :: CUDA Toolkit Documentation). If I add the -MP flag after -MMD I get an error:

nvcc fatal : Unknown option '-MP'

… I don’t see why it wouldn’t work, is this a WSL-specific problem?

2

Oh I have now seen that -MP was introduced to NVCC in CUDA Toolkit v11.1. The current version of the CUDA on WSL user guide (CUDA on WSL :: CUDA Toolkit Documentation) says to install CUA Toolkit v11.0

$ apt-get install -y cuda-toolkit-11-0

is this just a case of the user guide not being updated yet, and v11.1 is compatible with WSL?