Update to CMake 3.20 features, rapids-cmake and CPM (#187)

PR updates RAFT's CMake and uses CPM for dependency management. Still WIP, meant for early 0.20, but opening the PR to start debugging in CI. 

This is heavily based on cuDF's, RMM's and cumlprims' changes. PR does the following:

- [x] Adopt CPM:
    - [x] RMM
    - [x] FAISS
    - [x] GTest
    - [ ] NCCL: Missing building from source, will be added if/when required
    - [ ] UCX: Missing building from source, will be added if/when required
    - [x] CUB for CUDA < 11.0
- [x] Update arch detection 
- [x] Use generators to allow clang compilation as well
- [x] Updates to outdated cmake parts and remove old code
- [x] Update code to aid transition to #83

Authors:
  - Dante Gama Dessavre (https://github.com/dantegd)
  - Robert Maynard (https://github.com/robertmaynard)

Approvers:
  - Robert Maynard (https://github.com/robertmaynard)
  - Rick Ratzel (https://github.com/rlratzel)
  - Divye Gala (https://github.com/divyegala)

URL: #187

build.sh

@@ -42,7 +42,7 @@ PYTHON_DEPS_CLONE=${REPODIR}/python/external_repositories
 BUILD_DIRS="${CPP_RAFT_BUILD_DIR} ${PY_RAFT_BUILD_DIR} ${PYTHON_DEPS_CLONE}"
 
 # Set defaults for vars modified by flags to this script
-VERBOSE=""
+VERBOSE_FLAG=""
 BUILD_ALL_GPU_ARCH=0
 BUILD_GTEST=OFF
 BUILD_STATIC_FAISS=OFF
@@ -58,6 +58,9 @@ INSTALL_PREFIX=${INSTALL_PREFIX:=${PREFIX:=${CONDA_PREFIX}}}
 PARALLEL_LEVEL=${PARALLEL_LEVEL:=""}
 BUILD_ABI=${BUILD_ABI:=ON}
 
+# Default to Ninja if generator is not specified
+export CMAKE_GENERATOR="${CMAKE_GENERATOR:=Ninja}"
+
 function hasArg {
     (( ${NUMARGS} != 0 )) && (echo " ${ARGS} " | grep -q " $1 ")
 }
@@ -79,7 +82,8 @@ fi
 
 # Process flags
 if hasArg -v; then
-    VERBOSE=1
+    VERBOSE_FLAG=-v
+    set -x
 fi
 if hasArg -g; then
     BUILD_TYPE=Debug
@@ -130,41 +134,24 @@ fi
 # Configure for building all C++ targets
 if (( ${NUMARGS} == 0 )) || hasArg cppraft; then
     if (( ${BUILD_ALL_GPU_ARCH} == 0 )); then
-        GPU_ARCH=""
+        RAFT_CMAKE_CUDA_ARCHITECTURES="NATIVE"
         echo "Building for the architecture of the GPU in the system..."
     else
-        GPU_ARCH="-DGPU_ARCHS=ALL"
+        RAFT_CMAKE_CUDA_ARCHITECTURES="ALL"
         echo "Building for *ALL* supported GPU architectures..."
     fi
 
-    mkdir -p ${CPP_RAFT_BUILD_DIR}
-    cd ${CPP_RAFT_BUILD_DIR}
-
-    cmake -DNVTX=${NVTX} \
-          -DPARALLEL_LEVEL=${PARALLEL_LEVEL} \
-          -DNCCL_PATH=${INSTALL_PREFIX} \
+    cmake -S ${REPODIR}/cpp -B ${CPP_RAFT_BUILD_DIR} \
+          -DCMAKE_INSTALL_PREFIX=${INSTALL_PREFIX} \
+          -DCMAKE_CUDA_ARCHITECTURES=${RAFT_CMAKE_CUDA_ARCHITECTURES} \
+          -DNVTX=${NVTX} \
           -DDISABLE_DEPRECATION_WARNING=${BUILD_DISABLE_DEPRECATION_WARNING} \
           -DBUILD_GTEST=${BUILD_GTEST} \
-          -DBUILD_STATIC_FAISS=${BUILD_STATIC_FAISS} \
-          ..
-
-fi
-
-# Run all make targets at once
+          -DBUILD_STATIC_FAISS=${BUILD_STATIC_FAISS}
 
-MAKE_TARGETS=
-if hasArg cppraft; then
-    MAKE_TARGETS="${MAKE_TARGETS} test_raft"
-fi
-
-
-# If `./build.sh pyraft` is called, don't build C/C++ components
-if (( ${NUMARGS} == 0 )) || hasArg cppraft; then
-# If there are no targets specified when calling build.sh, it will
-# just call `make -j`. This avoids a lot of extra printing
-    cd ${CPP_RAFT_BUILD_DIR}
-    make -j${PARALLEL_LEVEL} ${MAKE_TARGETS} VERBOSE=${VERBOSE}
 
+    # Run all c++ targets at once
+    cmake --build  ${CPP_RAFT_BUILD_DIR} -j${PARALLEL_LEVEL} ${MAKE_TARGETS} ${VERBOSE_FLAG}
 fi