Rebuild for CUDA 12

.ci_support/linux_64_c_compiler_version10cuda_compiler_version11.2cxx_compiler_version10.yaml → .ci_support/linux_64_c_compiler_version10cuda_compilernvcccuda_compiler_version11.2cxx_compiler_version10.yaml

@@ -27,6 +27,7 @@ target_platform:
 zip_keys:
 - - c_compiler_version
   - cxx_compiler_version
+  - cuda_compiler
   - cuda_compiler_version
   - cdt_name
   - docker_image

.ci_support/linux_64_c_compiler_version12cuda_compilercuda-nvcccuda_compiler_version12.0cxx_compiler_version12.yaml

@@ -0,0 +1,33 @@
+c_compiler:
+- gcc
+c_compiler_version:
+- '12'
+cdt_name:
+- cos7
+channel_sources:
+- conda-forge
+channel_targets:
+- conda-forge main
+cuda_compiler:
+- cuda-nvcc
+cuda_compiler_version:
+- '12.0'
+cxx_compiler:
+- gxx
+cxx_compiler_version:
+- '12'
+docker_image:
+- quay.io/condaforge/linux-anvil-cos7-x86_64
+libblas:
+- 3.9 *netlib
+liblapack:
+- 3.9 *netlib
+target_platform:
+- linux-64
+zip_keys:
+- - c_compiler_version
+  - cxx_compiler_version
+  - cuda_compiler
+  - cuda_compiler_version
+  - cdt_name
+  - docker_image

.ci_support/linux_aarch64_c_compiler_version10cuda_compiler_version11.2cxx_compiler_version10.yaml → .ci_support/linux_aarch64_c_compiler_version10cuda_compilernvcccuda_compiler_version11.2cxx_compiler_version10.yaml

@@ -21,7 +21,7 @@ cxx_compiler:
 cxx_compiler_version:
 - '10'
 docker_image:
-- quay.io/condaforge/linux-anvil-aarch64-cuda:11.2
+- quay.io/condaforge/linux-anvil-cuda:11.2
 libblas:
 - 3.9 *netlib
 liblapack:
@@ -31,6 +31,7 @@ target_platform:
 zip_keys:
 - - c_compiler_version
   - cxx_compiler_version
+  - cuda_compiler
   - cuda_compiler_version
   - cdt_name
   - docker_image

.ci_support/migrations/cuda120.yaml

@@ -0,0 +1,80 @@
+migrator_ts: 1682985063
+__migrator:
+  kind:
+    version
+  migration_number:
+    1
+  build_number:
+    1
+  paused: false
+  override_cbc_keys:
+    - cuda_compiler_stub
+  operation: key_add
+  check_solvable: false
+  primary_key: cuda_compiler_version
+  ordering:
+    cxx_compiler_version:
+      - 9
+      - 8
+      - 7
+    c_compiler_version:
+      - 9
+      - 8
+      - 7
+    fortran_compiler_version:
+      - 9
+      - 8
+      - 7
+    docker_image:
+      - quay.io/condaforge/linux-anvil-comp7              # [os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-aarch64            # [os.environ.get("BUILD_PLATFORM") == "linux-aarch64"]
+      - quay.io/condaforge/linux-anvil-ppc64le            # [os.environ.get("BUILD_PLATFORM") == "linux-ppc64le"]
+      - quay.io/condaforge/linux-anvil-armv7l             # [os.environ.get("BUILD_PLATFORM") == "linux-armv7l"]
+      - quay.io/condaforge/linux-anvil-cuda:9.2           # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cuda:10.0          # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cuda:10.1          # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cuda:10.2          # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cuda:11.0          # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cuda:11.1          # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cuda:11.2          # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+      - quay.io/condaforge/linux-anvil-cos7-x86_64        # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64"]
+    cuda_compiler_version:
+      - None
+      - 10.2                       # [(linux64 or win) and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+      - 11.0                       # [(linux64 or win) and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+      - 11.1                       # [(linux64 or win) and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+      - 11.2                       # [(linux64 or win) and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+      - 12.0                       # [(linux64 or win) and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  commit_message: |
+    Rebuild for CUDA 12
+
+    The transition to CUDA 12 SDK includes new packages for all CUDA libraries and
+    build tools. Notably, the cudatoolkit package no longer exists, and packages
+    should depend directly on the specific CUDA libraries (libblas, libcusolver,
+    etc) as needed. For an in-depth overview of the changes and to report problems
+    [see this issue]( https://github.com/conda-forge/conda-forge.github.io/issues/1963 ).
+    Please feel free to raise any issues encountered there. Thank you! :pray:
+
+cuda_compiler:                 # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - cuda-nvcc                  # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+cuda_compiler_version:         # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - 12.0                       # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+c_compiler_version:            # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - 12                         # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+cxx_compiler_version:          # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - 12                         # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+fortran_compiler_version:      # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - 12                         # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+cudnn:                         # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - 8                          # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+cdt_name:                      # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - cos7                       # [linux64 and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+
+docker_image:                                      # [os.environ.get("BUILD_PLATFORM", "").startswith("linux-") and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]
+  - quay.io/condaforge/linux-anvil-cos7-x86_64     # [linux64 and os.environ.get("BUILD_PLATFORM") == "linux-64" and os.environ.get("CF_CUDA_ENABLED", "False") == "True"]

.ci_support/win_64_cuda_compiler_version11.2.yaml → .ci_support/win_64_cuda_compilernvcccuda_compiler_version11.2.yaml

@@ -16,3 +16,6 @@ liblapack:
 - 3.9 *netlib
 target_platform:
 - win-64
+zip_keys:
+- - cuda_compiler
+  - cuda_compiler_version

conda-forge.yml

@@ -1,15 +1,16 @@
+conda_build:
+  error_overlinking: true
+  pkg_format: '2'
 conda_forge_output_validation: true
 github:
   branch_name: main
   tooling_branch_name: main
-azure:
-  # Enable artifact storage to test with GPUs locally
-  store_build_artifacts: true
-conda_build:
-  pkg_format: '2'
 provider:
   linux_aarch64: default
   linux_ppc64le: default
 azure:
   # Enable artifact storage to test with GPUs locally
   store_build_artifacts: true
+build_platform:
+  linux_ppc64le: linux_64
+  linux_aarch64: linux_64

recipe/bld.bat

@@ -5,8 +5,8 @@ make generate --jobs %CPU_COUNT%
 if errorlevel 1 exit /b 1
 
 :: Duplicate lists because of https://bitbucket.org/icl/magma/pull-requests/32
-set "CUDA_ARCH_LIST=sm_35,sm_60,sm_70,sm_80"
-set "CUDAARCHS=35-virtual;60-virtual;70-virtual;80-virtual"
+set "CUDA_ARCH_LIST=sm_35,sm_50,sm_60,sm_70,sm_75,sm_80"
+set "CUDAARCHS=35-virtual;50-virtual;60-virtual;70-virtual;75-virtual;80-virtual"
 
 md build
 cd build

recipe/build.sh

@@ -3,13 +3,35 @@ set -exv
 # This step is required when building from raw source archive
 make generate --jobs ${CPU_COUNT}
 
+# Only about 7 virtual archs can be built 6 hours for CUDA 11
+
 # Duplicate lists because of https://bitbucket.org/icl/magma/pull-requests/32
-export CUDA_ARCH_LIST="sm_35,sm_50,sm_60,sm_70,sm_80"
-export CUDAARCHS="35-virtual;50-virtual;60-virtual;70-virtual;80-virtual"
+export CUDA_ARCH_LIST="sm_35,sm_50,sm_60,sm_61,sm_70,sm_75,sm_80"
+export CUDAARCHS="35-virtual;50-virtual;60-virtual;61-virtual;70-virtual;75-virtual;80-virtual"
+
+if [[ "$cuda_compiler_version" == "12.0" ]]; then
+  export CUDA_ARCH_LIST="sm_50,sm_60,sm_61,sm_70,sm_75,sm_80,sm_86,sm_89,sm_90"
+  export CUDAARCHS="50-virtual;60-virtual;61-virtual;70-virtual;75-virtual;80-virtual;86-virtual;89-virtual;90-virtual"
+fi
+
+if [[ "$target_platform" == "linux-ppc64le" ]]; then
+  export CUDA_ARCH_LIST="sm_50,sm_60,sm_61,sm_70,sm_75,sm_80,sm_86"
+  export CUDAARCHS="50-virtual;60-virtual;61-virtual;70-virtual;75-virtual;80-virtual;86-virtual"
+fi
+
+# Jetsons are more common for ARM devices, so target those minor versions
+if [[ "$target_platform" == "linux-aarch64" ]]; then
+  export CUDA_ARCH_LIST="sm_50,sm_53,sm_60,sm_62,sm_70,sm_72,sm_80,"
+  export CUDAARCHS="50-virtual;53-virtual;60-virtual;62-virtual;70-virtual;72-virtual;80-virtual"
+fi
 
 # Remove CXX standard flags added by conda-forge. std=c++11 is required to
 # compile some .cu files
-CXXFLAGS="${CXXFLAGS//-std=c++17/-std=c++11}"
+export CXXFLAGS="${CXXFLAGS//-std=c++17/-std=c++11}"
+
+# Conda-forge nvcc compiler flags environment variable doesn't match CMake environment variable
+# Redirect it so that the flags are added to nvcc calls
+export CUDAFLAGS="${CUDAFLAGS} ${CUDA_CFLAGS}"
 
 mkdir build
 cd build

recipe/meta.yaml

@@ -84,6 +84,8 @@
   'z_v1_map'
 ] %}
 
+{% set cuda_major = environ.get("cuda_compiler_version", "11.2").split(".")[0]|int %}
+
 package:
   name: libmagma_sparse
   version: {{ version }}
@@ -99,7 +101,7 @@ source:
       - manual-link.patch
 
 build:
-  number: 2
+  number: 3
   skip: true  # [cuda_compiler_version == "None"]
   skip: true  # [cuda_compiler_version == "10.2"]
   skip: true  # [cuda_compiler_version == "11.0"]
@@ -124,6 +126,12 @@ requirements:
     - ninja
     - python
   host:
+{% if cuda_major == 12 %}
+    - cuda-cudart-dev
+    - cuda-profiler-api
+    - libcublas-dev
+    - libcusparse-dev
+{% endif %}
     - liblapack
     - libblas
     - libmagma {{ version }}