Replies: 1 comment
-
The DPC++ build assumes that ROCm is installed in /opt/rocm, if it is installed somewhere else, the directory must be provided through the CMake variable SYCL_BUILD_PI_HIP_ROCM_DIR which can be passed using the --cmake-opt option of configure.py as follows: python $DPCPP_HOME/llvm/buildbot/configure.py --hip |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
My system installs ROCm to locations unknown to the DPC++ build system, so when building DPC++, I need to use
DSYCL_BUILD_PI_HIP_ROCM_DIR
,DSYCL_BUILD_PI_HIP_INCLUDE_DIR
,DSYCL_BUILD_PI_HIP_HSA_INCLUDE_DIR
, andDSYCL_BUILD_PI_HIP_LIB_DIR
. I was able to complete the build.However, after installing, targeting HIP ROCm creates the following error message:
I was able to solve the problem by passing
--rocm-device-lib-path=/usr/lib/amdgcn/bitcode/
manually. But is there a way to change this search path at build time? Which build-time variable is controlling it?Beta Was this translation helpful? Give feedback.
All reactions