Rebuild for CUDA 12

[regro-cf-autotick-bot]

This PR has been triggered in an effort to update cuda120.

Notes and instructions for merging this PR:

1. Please merge the PR only after the tests have passed.
2. Feel free to push to the bot's branch to update this PR if needed.

Please note that if you close this PR we presume that the feedstock has been rebuilt, so if you are going to perform the rebuild yourself don't close this PR until the your rebuild has been merged.

If this PR was opened in error or needs to be updated please add the bot-rerun label to this PR. The bot will close this PR and schedule another one. If you do not have permissions to add this label, you can use the phrase @conda-forge-admin, please rerun bot in a PR comment to have the conda-forge-admin add it for you.

_{This PR was created by the regro-cf-autotick-bot. The regro-cf-autotick-bot is a service to automatically track the dependency graph, migrate packages, and propose package version updates for conda-forge. Feel free to drop us a line if there are any issues! This PR was generated by https://github.com/regro/cf-scripts/actions/runs/5139371143, please use this URL for debugging.}

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[jakirkham]

Seeing the following error on CI:

CMake Error at cmake/CMakeHelper.cmake:126 (target_link_libraries):
  Target "pba" links to:

    CUDA::curand

As the new CUDA 12 packages split all these components up ( conda-forge/staged-recipes#21382 ), think we need to add libcurand-dev to host:

  host:
    - libcurand-dev      # [(cuda_compiler_version or "").startswith("12")]

Note that *-dev suffixed packages have headers and dynamic libraries. They add the * package (no -dev suffix) to run automatically via run_exports

Edit: For more info about the CUDA 12 bringup, please see issue ( conda-forge/conda-forge.github.io#1963 )

[Tobias-Fischer]

It seems that upstream does not yet support cuda 12. It uses some deprecated API.

[Tobias-Fischer]

See colmap/colmap#1840

[jakirkham]

Gotcha. Thanks for investigating. Going to mark this as draft then. Can mark as ready to review once that changes

[Tobias-Fischer]

Hi @jakirkham - could you please have another look?

I'm now stuck with:

[ 19%] Building CUDA object src/CMakeFiles/colmap_cuda.dir/mvs/gpu_mat_ref_image.cu.o
In file included from /home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/cuda_array_wrapper.h:39,
                 from /home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/gpu_mat_ref_image.h:37,
                 from /home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/gpu_mat_ref_image.cu:32:
/home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/gpu_mat.h:41:10: fatal error: curand_kernel.h: No such file or directory
   41 | #include <curand_kernel.h>
      |          ^~~~~~~~~~~~~~~~~
compilation terminated.
make[2]: *** [src/CMakeFiles/colmap_cuda.dir/build.make:92: src/CMakeFiles/colmap_cuda.dir/mvs/gpu_mat_ref_image.cu.o] Error 1
make[2]: *** Waiting for unfinished jobs....
In file included from /home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/gpu_mat_prng.h:35,
                 from /home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/gpu_mat_prng.cu:32:
/home/conda/feedstock_root/build_artifacts/colmap_1685723949673/work/src/mvs/gpu_mat.h:41:10: fatal error: curand_kernel.h: No such file or directory
   41 | #include <curand_kernel.h>
      |          ^~~~~~~~~~~~~~~~~
compilation terminated.
make[2]: *** [src/CMakeFiles/colmap_cuda.dir/build.make:77: src/CMakeFiles/colmap_cuda.dir/mvs/gpu_mat_prng.cu.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:544: src/CMakeFiles/colmap_cuda.dir/all] Error 2

[Tobias-Fischer]

Kind ping @jakirkham :)

[jakirkham]

Oops didn't even see the first ping 😅

We just merged this fix from Daniel ( conda-forge/cuda-nvcc-feedstock#19 ), which may help here

The packages are already up (though maybe still mirroring to CDN). Would try restarting CI in maybe ~30mins or so

Please let me know how that goes 🙂

[jakirkham]

Looks like we are seeing a new issue in CI:

[ 19%] Building CUDA object src/CMakeFiles/colmap_cuda.dir/mvs/gpu_mat_prng.cu.o
/home/conda/feedstock_root/build_artifacts/colmap_1686132516870/work/src/mvs/gpu_mat_ref_image.cu(42): error: texture is not a template

/home/conda/feedstock_root/build_artifacts/colmap_1686132516870/work/src/mvs/gpu_mat_ref_image.cu(57): error: no instance of overloaded function "tex2D" matches the argument list
            argument types are: (<error-type>, const size_t, const size_t)

/home/conda/feedstock_root/build_artifacts/colmap_1686132516870/work/src/mvs/gpu_mat_ref_image.cu(68): error: no instance of overloaded function "tex2D" matches the argument list
            argument types are: (<error-type>, unsigned long, unsigned long)

/home/conda/feedstock_root/build_artifacts/colmap_1686132516870/work/src/mvs/gpu_mat_ref_image.cu(110): error: identifier "cudaBindTextureToArray" is undefined

/home/conda/feedstock_root/build_artifacts/colmap_1686132516870/work/src/mvs/gpu_mat_ref_image.cu(115): error: identifier "cudaUnbindTexture" is undefined

5 errors detected in the compilation of "/home/conda/feedstock_root/build_artifacts/colmap_1686132516870/work/src/mvs/gpu_mat_ref_image.cu".

[github-actions]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• azure: passed

Thus the PR was passing and merged! Have a great day!