[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: CMake musings for ROCm 4.5.2 (Was Re: ROCm packaging session memo) (Was Re: ROCm installation)



Hello,

Thank you! This looks spot on!

BR, Maxime

On 2/4/22 01:58, Cordell Bloor wrote:
Hi Maxime,

On 2022-01-30 16:59, Maxime Chambonnet wrote:
For rocSOLVER, these four lines seem to be what is crippling the
underlying install function:
rocsolver/library/src/CMakeLists.txt, line 249

rocm_install_targets(
  TARGETS ${package_targets}
*********************************************************
  INCLUDE
    ${CMAKE_SOURCE_DIR}/library/include
    ${CMAKE_BINARY_DIR}/include
  PREFIX rocsolver
*********************************************************
)

I suspect that without them, we get very close to "Debian's preferred
install layout"!

And with PARSE_PREFIX and INSTALL_PREFIX, you still have latitude to
install wherever spack wants, or in /opt/rocm?

The current proposal is deleting those lines rocm_install_targets, the PREFIX from rocm_export_targets and removing rocm_install_symlink_subdirs entirely, in every ROCm math library. (There's additional complexity for maintaining backwards-compatibility with the old /opt/rocm layout, but it's gated behind a CMake flag and presumably irrelevant to Debian.)

I'm reviewing the change that is supposed to fix this for you in rocBLAS. When I have a moment, I'll port equivalent changes to rocSOLVER in 4.5 and update here so you can take a look.

Sincerely,
Cory Bloor



Reply to: