Rebuild cuda120 0 2 hdbc077

RaulPPelaez commented 11 months ago

Checklist

[x] Used a personal fork of the feedstock to propose changes
[x] Bumped the build number (if the version is unchanged)
[ ] Reset the build number to 0 (if the version changed)
[x] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
[x] Ensured the license file is being packaged.

This is just #192 but I added the archs that prevent CUDA 12 builds from starting the compilation.

conda-forge-webservices[bot] commented 11 months ago

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.

RaulPPelaez commented 11 months ago

@conda-forge-admin, please rerender

RaulPPelaez commented 11 months ago

CUDA libraries are not found:

  --     CUDA root directory : /home/conda/feedstock_root/build_artifacts/pytorch-recipe_1698233734210/_build_env
  --     CUDA library        : CUDA_CUDA_LIB-NOTFOUND
  --     cudart library      : /home/conda/feedstock_root/build_artifacts/pytorch-recipe_1698233734210/_build_env/targets/x86_64-linux/lib/libcudart.so
  --     cublas library      : CUDA_cublas_LIBRARY-NOTFOUND;CUDA_cublasLt_LIBRARY-NOTFOUND
  --     cufft library       : CUDA_cufft_LIBRARY-NOTFOUND
  --     curand library      : CUDA_curand_LIBRARY-NOTFOUND
  --     cuDNN library       : /home/conda/feedstock_root/build_artifacts/pytorch-recipe_1698233734210/_h_env_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placehold_placeh/lib/libcudnn.so
  --     nvrtc               : CUDA_NVRTC_LIB-NOTFOUND
  --     CUDA include path   : /home/conda/feedstock_root/build_artifacts/pytorch-recipe_1698233734210/_build_env/targets/x86_64-linux/include
  --     NVCC executable     : /home/conda/feedstock_root/build_artifacts/pytorch-recipe_1698233734210/_build_env/bin/nvcc
  --     CUDA compiler       : /home/conda/feedstock_root/build_artifacts/pytorch-recipe_1698233734210/_build_env/bin/nvcc
  --     CUDA flags          :  -Xfatbin -compress-all -DONNX_NAMESPACE=onnx_torch -gencode arch=compute_50,code=sm_50 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_89,code=sm_89 -gencode arch=compute_90,code=sm_90 -gencode arch=compute_90,code=compute_90 -Xcudafe --diag_suppress=cc_clobber_ignored,--diag_suppress=integer_sign_change,--diag_suppress=useless_using_declaration,--diag_suppress=set_but_not_used,--diag_suppress=field_without_dll_interface,--diag_suppress=base_class_has_different_dll_interface,--diag_suppress=dll_interface_conflict_none_assumed,--diag_suppress=dll_interface_conflict_dllexport_assumed,--diag_suppress=implicit_return_from_non_void_function,--diag_suppress=unsigned_compare_with_zero,--diag_suppress=declared_but_not_referenced,--diag_suppress=bad_friend_decl --expt-relaxed-constexpr --expt-extended-lambda -Xfatbin -compress-all -Wno-deprecated-gpu-targets --expt-extended-lambda -DCUB_WRAPPED_NAMESPACE=at_cuda_detail -DCUDA_HAS_FP16=1 -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__

These are the installed packages:

The following NEW packages will be INSTALLED:

    _libgcc_mutex:        0.1-conda_forge             conda-forge
    _openmp_mutex:        4.5-2_kmp_llvm              conda-forge
    brotli-python:        1.1.0-py310hc6cd4ac_1       conda-forge
    bzip2:                1.0.8-h7f98852_4            conda-forge
    ca-certificates:      2023.7.22-hbcca054_0        conda-forge
    certifi:              2023.7.22-pyhd8ed1ab_0      conda-forge
    charset-normalizer:   3.3.1-pyhd8ed1ab_0          conda-forge
    cuda-cudart:          12.0.107-h59595ed_6         conda-forge
    cuda-cudart_linux-64: 12.0.107-h59595ed_6         conda-forge
    cuda-nvrtc:           12.0.76-h59595ed_1          conda-forge
    cuda-version:         12.0-hffde075_2             conda-forge
    cudnn:                8.8.0.121-h264754d_3        conda-forge
    future:               0.18.3-pyhd8ed1ab_0         conda-forge
    icu:                  73.2-h59595ed_0             conda-forge
    idna:                 3.4-pyhd8ed1ab_0            conda-forge
    ld_impl_linux-64:     2.40-h41732ed_0             conda-forge
    libabseil:            20230802.1-cxx17_h59595ed_0 conda-forge
    libblas:              3.9.0-16_linux64_mkl        conda-forge
    libcblas:             3.9.0-16_linux64_mkl        conda-forge
    libcublas:            12.0.1.189-hcb278e6_2       conda-forge
    libcusparse:          12.0.0.76-hcb278e6_1        conda-forge
    libffi:               3.4.2-h7f98852_5            conda-forge
    wheel:                0.41.2-pyhd8ed1ab_0         conda-forge
    xz:                   5.2.6-h166bdaf_0            conda-forge
    yaml:                 0.2.5-h7f98852_2            conda-forge
    zstd:                 1.5.5-hfc55251_0            conda-forge

Should "{{ compiler('cuda') }}" bring the dev versions of the CUDA libraries? @mikemhenry could you please offer some guidance here?

conda-forge-webservices[bot] commented 11 months ago

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

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

Failed to even lint the recipe, probably because of a conda-smithy bug :cry:. This likely indicates a problem in your meta.yaml, though. To get a traceback to help figure out what's going on, install conda-smithy and run conda smithy recipe-lint . from the recipe directory.

conda-forge-webservices[bot] commented 11 months ago

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 commented 11 months ago

Thanks Raul! 🙏

Added a small suggestion above

We could also look at moving the export CUDA_TOOLKIT_ROOT_DIR=$CUDA_HOME line into the cuda_compiler_version branches. With CUDA 11 and earlier, this logic is needed since the CTK install lived outside Conda. With CUDA 12, the CTK is now within Conda so we stopped setting $CUDA_HOME. AFAICT PyTorch doesn't mind this and just ignores CUDA_TOOLKIT_ROOT_DIR, which is also fine. Just wanted to mention it for transparency

Would double check all of the libraries are needed. Otherwise this looks good. Seems to build pretty far (ofc over CI time limits, but that's a different story)

@hmaarrfk what do you think? 🙂

hmaarrfk commented 11 months ago

please move what you need now. I'm going to add the proto if migration tonight and restart a build process. takes about 48 hours....

conda-forge / pytorch-cpu-feedstock

Rebuild cuda120 0 2 hdbc077 #193