various updates to take suggestions into account

- `EESSI-install-software.sh`
  - use `scripts/gpu_support/nvidia/install_cuda_and_libraries.sh` with
    `scripts/gpu_support/nvidia/eessi-2023.06-cuda-and-libraries.yml`
- `create_lmodsitepackage.py`
  - consolidate `eessi_{cuda,cudnn}_enabled_load_hook` functions in a single one
    (`eessi_cuda_and_libraries_enabled_load_hook`)
  - the remaining hook is prepared to easily add new modules, e.g., cuTENSOR
- `eb_hooks.py`
  - put code that iterates over all files replacing non-distributable ones with
    symlinks into `host_injections` with a common function
    (`replace_non_distributable_files_with_symlinks`)
- `install_scripts.sh`
  - add files to copy to CVMFS (see `nvidia_files`)
- `scripts/gpu_support/nvidia/install_cuda_and_libraries.sh`
  - improved creation of tmp directory

EESSI-install-software.sh

@@ -233,8 +233,10 @@ else
 fi
 
 if [ -z "${skip_cuda_install}" ] || [ ! "${skip_cuda_install}" ]; then
-    ${EESSI_PREFIX}/scripts/gpu_support/nvidia/install_cuda_host_injections.sh -c 12.1.1 --accept-cuda-eula
-    ${EESSI_PREFIX}/scripts/gpu_support/nvidia/install_cudnn_host_injections.sh -c 12.1.1 -d 8.9.2.26
+    ${EESSI_PREFIX}/scripts/gpu_support/nvidia/install_cuda_and_libraries.sh \
+        -e ${EESSI_PREFIX}/scripts/gpu_support/nvidia/eessi-2023.06-cuda-and-libraries.yml \
+        -t /tmp/temp \
+        --accept-cuda-eula
 else
     echo "Skipping installation of CUDA SDK and cu* libraries in host_injections, since the --skip-cuda-install flag was passed OR no EasyBuild module was found"
 fi

create_lmodsitepackage.py

@@ -107,35 +107,41 @@
 end
 
 
-local function eessi_cuda_enabled_load_hook(t)
+local function eessi_cuda_and_libraries_enabled_load_hook(t)
     local frameStk  = require("FrameStk"):singleton()
     local mt        = frameStk:mt()
     local simpleName = string.match(t.modFullName, "(.-)/")
-    -- If we try to load CUDA itself, check if the full CUDA SDK was installed on the host in host_injections. 
-    -- This is required for end users to build additional CUDA software. If the full SDK isn't present, refuse
-    -- to load the CUDA module and print an informative message on how to set up GPU support for EESSI
+    local packagesList = { ["CUDA"] = true, ["cuDNN"] = true }
+    -- If we try to load any of the modules in packagesList, we check if the
+    -- full package was installed on the host in host_injections.
+    -- This is required for end users to build additional software that depends
+    -- on the package. If the full SDK isn't present, refuse
+    -- to load the module and print an informative message on how to set up GPU support for EESSI
     local refer_to_docs = "For more information on how to do this, see https://www.eessi.io/docs/gpu/.\\n"
-    if simpleName == 'CUDA' then
+    if packagesList[simpleName] then
+        -- simpleName is a module in packagesList
         -- get the full host_injections path
         local hostInjections = string.gsub(os.getenv('EESSI_SOFTWARE_PATH') or "", 'versions', 'host_injections')
-        -- build final path where the CUDA software should be installed
-        local cudaEasyBuildDir = hostInjections .. "/software/" .. t.modFullName .. "/easybuild"
-        local cudaDirExists = isDir(cudaEasyBuildDir)
-        if not cudaDirExists then
+
+        -- build final path where the software should be installed
+        local packageEasyBuildDir = hostInjections .. "/software/" .. t.modFullName .. "/easybuild"
+        local packageDirExists = isDir(packageEasyBuildDir)
+        if not packageDirExists then
             local advice = "but while the module file exists, the actual software is not entirely shipped with EESSI "
-            advice = advice .. "due to licencing. You will need to install a full copy of the CUDA SDK where EESSI "
+            advice = advice .. "due to licencing. You will need to install a full copy of the " .. simpleName .. " package where EESSI "
             advice = advice .. "can find it.\\n"
             advice = advice .. refer_to_docs
             LmodError("\\nYou requested to load ", simpleName, " ", advice)
         end
     end
-    -- when loading CUDA enabled modules check if the necessary driver libraries are accessible to the EESSI linker,
+    -- when loading CUDA (and cu*) enabled modules check if the necessary driver libraries are accessible to the EESSI linker,
     -- otherwise, refuse to load the requested module and print error message
     local haveGpu = mt:haveProperty(simpleName,"arch","gpu")
     if haveGpu then
         local arch = os.getenv("EESSI_CPU_FAMILY") or ""
-        local cudaVersionFile = "/cvmfs/software.eessi.io/host_injections/nvidia/" .. arch .. "/latest/cuda_version.txt"
-        local cudaDriverFile = "/cvmfs/software.eessi.io/host_injections/nvidia/" .. arch .. "/latest/libcuda.so"
+        local cvmfs_repo = os.getenv("EESSI_CVMFS_REPO") or ""
+        local cudaVersionFile = cvmfs_repo .. "/host_injections/nvidia/" .. arch .. "/latest/cuda_version.txt"
+        local cudaDriverFile = cvmfs_repo .. "/host_injections/nvidia/" .. arch .. "/latest/libcuda.so"
         local cudaDriverExists = isFile(cudaDriverFile)
         local singularityCudaExists = isFile("/.singularity.d/libs/libcuda.so")
         if not (cudaDriverExists or singularityCudaExists)  then
@@ -172,38 +178,13 @@
     end
 end
 
-local function eessi_cudnn_enabled_load_hook(t)
-    local frameStk  = require("FrameStk"):singleton()
-    local mt        = frameStk:mt()
-    local simpleName = string.match(t.modFullName, "(.-)/")
-    -- If we try to load cuDNN itself, check if the full cuDNN package was installed on the host in host_injections. 
-    -- This is required for end users to build additional cuDNN dependent software. If the full SDK isn't present, refuse
-    -- to load the cuDNN module and print an informative message on how to set up GPU support for EESSI
-    local refer_to_docs = "For more information on how to do this, see https://www.eessi.io/docs/gpu/.\\n"
-    if simpleName == 'cuDNN' then
-        -- get the full host_injections path
-        local hostInjections = string.gsub(os.getenv('EESSI_SOFTWARE_PATH') or "", 'versions', 'host_injections')
-        -- build final path where the cuDNN software should be installed
-        local cudnnEasyBuildDir = hostInjections .. "/software/" .. t.modFullName .. "/easybuild"
-        local cudnnDirExists = isDir(cudnnEasyBuildDir)
-        if not cudnnDirExists then
-            local advice = "but while the module file exists, the actual software is not entirely shipped with EESSI "
-            advice = advice .. "due to licencing. You will need to install a full copy of the cuDNN package where EESSI "
-            advice = advice .. "can find it.\\n"
-            advice = advice .. refer_to_docs
-            LmodError("\\nYou requested to load ", simpleName, " ", advice)
-        end
-    end
-end
-
 -- Combine both functions into a single one, as we can only register one function as load hook in lmod
 -- Also: make it non-local, so it can be imported and extended by other lmodrc files if needed
 function eessi_load_hook(t)
     -- Only apply CUDA and cuDNN hooks if the loaded module is in the EESSI prefix
     -- This avoids getting an Lmod Error when trying to load a CUDA and cuDNN module from a local software stack
     if from_eessi_prefix(t) then
-        eessi_cuda_enabled_load_hook(t)
-        eessi_cudnn_enabled_load_hook(t)
+        eessi_cuda_and_libraries_enabled_load_hook(t)
     end
 end
 

eb_hooks.py

@@ -567,6 +567,47 @@ def post_sanitycheck_hook(self, *args, **kwargs):
         POST_SANITYCHECK_HOOKS[self.name](self, *args, **kwargs)
 
 
+def replace_non_distributable_files_with_symlinks(log, install_dir, package, allowlist):
+    """
+    Replace files that cannot be distributed with symlinks into host_injections
+    """
+    extension_based = { "CUDA": False, "cuDNN": True }
+    if not package in extension_based:
+        raise EasyBuildError("Don't know how to strip non-distributable files from package %s.", package)
+
+    # iterate over all files in the package installation directory
+    for dir_path, _, files in os.walk(install_dir):
+        for filename in files:
+            full_path = os.path.join(dir_path, filename)
+            # we only really care about real files, i.e. not symlinks
+            if not os.path.islink(full_path):
+                # check if the current file name stub is part of the allowlist
+                basename =  filename.split('.')[0]
+                if extension_based[package]:
+                    if '.' in filename:
+                        extension = '.' + filename.split('.')[1]
+                if basename in allowlist:
+                    log.debug("%s is found in allowlist, so keeping it: %s", basename, full_path)
+                elif extension_based[package] and '.' in filename and extension in allowlist:
+                    log.debug("%s is found in allowlist, so keeping it: %s", extension, full_path)
+                else:
+                    if extension_based[package]:
+                        print_name = filename
+                    else:
+                        print_name = basename
+                    log.debug("%s is not found in allowlist, so replacing it with symlink: %s",
+                              print_name, full_path)
+                    # if it is not in the allowlist, delete the file and create a symlink to host_injections
+                    host_inj_path = full_path.replace('versions', 'host_injections')
+                    # make sure source and target of symlink are not the same
+                    if full_path == host_inj_path:
+                        raise EasyBuildError("Source (%s) and target (%s) are the same location, are you sure you "
+                                             "are using this hook for an EESSI installation?",
+                                             full_path, host_inj_path)
+                    remove_file(full_path)
+                    symlink(host_inj_path, full_path)
+
+
 def post_sanitycheck_cuda(self, *args, **kwargs):
     """
     Remove files from CUDA installation that we are not allowed to ship,
@@ -606,28 +647,9 @@ def post_sanitycheck_cuda(self, *args, **kwargs):
         if 'libcudart' not in allowlist:
             raise EasyBuildError("Did not find 'libcudart' in allowlist: %s" % allowlist)
 
-        # iterate over all files in the CUDA installation directory
-        for dir_path, _, files in os.walk(self.installdir):
-            for filename in files:
-                full_path = os.path.join(dir_path, filename)
-                # we only really care about real files, i.e. not symlinks
-                if not os.path.islink(full_path):
-                    # check if the current file name stub is part of the allowlist
-                    basename = filename.split('.')[0]
-                    if basename in allowlist:
-                        self.log.debug("%s is found in allowlist, so keeping it: %s", basename, full_path)
-                    else:
-                        self.log.debug("%s is not found in allowlist, so replacing it with symlink: %s",
-                                       basename, full_path)
-                        # if it is not in the allowlist, delete the file and create a symlink to host_injections
-                        host_inj_path = full_path.replace('versions', 'host_injections')
-                        # make sure source and target of symlink are not the same
-                        if full_path == host_inj_path:
-                            raise EasyBuildError("Source (%s) and target (%s) are the same location, are you sure you "
-                                                 "are using this hook for an EESSI installation?",
-                                                 full_path, host_inj_path)
-                        remove_file(full_path)
-                        symlink(host_inj_path, full_path)
+        # replace files that are not distributable with symlinks into
+        # host_injections
+        replace_non_distributable_files_with_symlinks(self.log, self.installdir, self.name, allowlist)
     else:
         raise EasyBuildError("CUDA-specific hook triggered for non-CUDA easyconfig?!")
 
@@ -643,8 +665,7 @@ def post_sanitycheck_cudnn(self, *args, **kwargs):
 
         allowlist = ['LICENSE']
 
-        # read cuDNN LICENSE, construct allowlist based on section 2. Distribution
-        # that specifies list of files that can be shipped
+        # read cuDNN LICENSE, construct allowlist based on section "2. Distribution" that specifies list of files that can be shipped
         license_path = os.path.join(self.installdir, 'LICENSE')
         search_string = "2. Distribution. The following portions of the SDK are distributable under the Agreement:"
         with open(license_path) as infile:
@@ -660,32 +681,9 @@ def post_sanitycheck_cudnn(self, *args, **kwargs):
         allowlist = sorted(set(allowlist))
         self.log.info("Allowlist for files in cuDNN installation that can be redistributed: " + ', '.join(allowlist))
 
-        # iterate over all files in the cuDNN installation directory
-        for dir_path, _, files in os.walk(self.installdir):
-            for filename in files:
-                full_path = os.path.join(dir_path, filename)
-                # we only really care about real files, i.e. not symlinks
-                if not os.path.islink(full_path):
-                    # check if the current file is part of the allowlist
-                    basename = filename.split('.')[0]
-                    if '.' in filename:
-                        extension = '.' + filename.split('.')[1]
-                    if basename in allowlist:
-                        self.log.debug("%s is found in allowlist, so keeping it: %s", basename, full_path)
-                    elif '.' in filename and extension in allowlist:
-                        self.log.debug("%s is found in allowlist, so keeping it: %s", extension, full_path)
-                    else:
-                        self.log.debug("%s is not found in allowlist, so replacing it with symlink: %s",
-                                       filename, full_path)
-                        # if it is not in the allowlist, delete the file and create a symlink to host_injections
-                        host_inj_path = full_path.replace('versions', 'host_injections')
-                        # make sure source and target of symlink are not the same
-                        if full_path == host_inj_path:
-                            raise EasyBuildError("Source (%s) and target (%s) are the same location, are you sure you "
-                                                 "are using this hook for a EESSI installation?",
-                                                 full_path, host_inj_path)
-                        remove_file(full_path)
-                        symlink(host_inj_path, full_path)
+        # replace files that are not distributable with symlinks into
+        # host_injections
+        replace_non_distributable_files_with_symlinks(self.log, self.installdir, self.name, allowlist)
     else:
         raise EasyBuildError("cuDNN-specific hook triggered for non-cuDNN easyconfig?!")
 

install_scripts.sh

@@ -110,7 +110,11 @@ copy_files_by_list ${TOPDIR}/scripts ${INSTALL_PREFIX}/scripts "${script_files[@
 
 # Copy files for the scripts/gpu_support/nvidia directory
 nvidia_files=(
-    install_cuda_host_injections.sh install_cudnn_host_injections.sh link_nvidia_host_libraries.sh
+    eessi-2023.06-cuda-and-libraries.yml
+    install_cuda_and_libraries.sh
+    install_cuda_host_injections.sh
+    install_cudnn_host_injections.sh
+    link_nvidia_host_libraries.sh
 )
 copy_files_by_list ${TOPDIR}/scripts/gpu_support/nvidia ${INSTALL_PREFIX}/scripts/gpu_support/nvidia "${nvidia_files[@]}"
 

scripts/gpu_support/nvidia/install_cuda_and_libraries.sh

@@ -93,11 +93,13 @@ export EESSI_SITE_INSTALL=${EESSI_SOFTWARE_PATH/versions/host_injections}
 if [[ -z "${TEMP_DIR}" ]]; then
     tmpdir=$(mktemp -d)
 else
-    tmpdir="${TEMP_DIR}"/temp
-    if ! mkdir "$tmpdir" ; then
+    mkdir -p ${TEMP_DIR}
+    tmpdir=$(mktemp -d --tmpdir=${TEMP_DIR} cuda_n_co.XXX)
+    if [[ ! -d "$tmpdir" ]] ; then
         fatal_error "Could not create directory ${tmpdir}"
     fi
 fi
+echo "Created temporary directory '${tmpdir}'"
 
 # workaround for EasyBuild not being found when loading "extend" module
 module load EasyBuild/4.9.1