CLion - CUDA Syntax - "Use of undeclared identifier cudaConfigureCall"

Answered

I have a simple HelloWorld CUDA application that builds and runs correctly. I'm using CLion 2020.2 which supports CUDA syntax however I'm getting a false error "Use of undeclared identifier cudaConfigureCall". According to this JetBrains site CLion should support this syntax.

I'm building this CMake project on Ubuntu 20.04 using CUDA 11.0 compiled with nvcc, gcc-8.0 and g++-8.0

Any thoughts?

 

Here's my CMakeLists.txt and Toolchain settings

cmake_minimum_required(VERSION 3.16)
project(LinuxCudaTest CUDA CXX)

set(CMAKE_CUDA_STANDARD 14)

add_executable(LinuxCudaTest main.cu)

set_target_properties(
LinuxCudaTest
PROPERTIES
CUDA_SEPARABLE_COMPILATION ON)

2 comments
Comment actions Permalink

I see the same, on arch linux using cuda 11.1.1-1 and gcc 10.2.0-3 

0

Please sign in to leave a comment.