[Docs] Add Documentation on BOLT Build Configs

[aidengrossman]

Summary

This patch adds documentation into the advanced builds documentation on
how to use the BOLT caches, including the combinations with the PGO
multistage builds and (Thin)LTO.

[aidengrossman]

Adds info on using BOLT optimized builds and an example configuration using PGO, ThinLTO, and BOLT for a highly optimized build. Really compelling performance gains were demonstrated at the LLVM US Developer's Meeting, and I wanted to make replication more accessible through better documentation. It also seems like these details are also documented in clang/cmake/caches/README.txt. If there's interest for adding documentation there as well, I can open up another patch documenting some of this stuff in more detail (ie enumerating targets added by the BOLT caches as done in the other targets), but the CMake cache files are relatively simple.

These configs are mostly taken from the commit messages where they are implemented, with some slight modifications, particularly removing the LLVM_ENABLE_LLD flag in the fully optimized build config in case the user doesn't have LLD installed. I've left the other LLD flags on due to the significantly faster link time due to using LLD.

[sylvestre.ledru]

llvm/docs/AdvancedBuilds.rst

201

Can you add a link to the bolt github readme here ?

[Amir]

Thank you for adding the documentation! Looks good to me, clear and concise.

[aidengrossman]

Added link to BOLT readme within the section on BOLT
documentation.

llvm/docs/AdvancedBuilds.rst

 
 **stage2-check-all**
   Depends on stage2 and runs check-all using the stage2 compiler.
 
 **stage2-test-suite**
   Depends on stage2 and runs the test-suite using the stage2 compiler (requires
   in-tree test-suite).
 
+BOLT
+====
+
+`BOLT <https://github.com/llvm/llvm-project/blob/main/bolt/README.md>`_
+(Binary Optimization and Layout Tool) is a tool that optimizes binaries
+post-link by profiling them at runtime and then using that information to
+optimize the layout of the final binary among other optimizations performed
+at the binary level. There are also CMake caches available to build
+LLVM/Clang with BOLT.
+
+To configure a single-stage build that builds LLVM/Clang and then optimizes
+it with BOLT, use the following CMake configuration:
+
+.. code-block:: console
+
+  $ cmake <path to source>/llvm -C <path to source>/clang/cmake/caches/BOLT.cmake
+
+Then, build the BOLT-optimized binary by running the following ninja command:
+
+.. code-block:: console
+
+  $ ninja clang++-bolt
+
+If you're seeing errors in the build process, try building with a recent
+version of Clang/LLVM by setting the CMAKE_C_COMPILER and
+CMAKE_CXX_COMPILER flags to the appropriate values.
+
+It is also possible to use BOLT on top of PGO and (Thin)LTO for an even more
+significant runtime speedup. To configure a three stage PGO build with ThinLTO
+that optimizes the resulting binary with BOLT, use the following CMake
+configuration command:
+
+.. code-block:: console
+
+  $ cmake -G Ninja <path to source>/llvm \
+      -C <path to source>/clang/cmake/caches/BOLT-PGO.cmake \
+      -DBOOTSTRAP_LLVM_ENABLE_LLD=ON \
+      -DBOOTSTRAP_BOOTSTRAP_LLVM_ENABLE_LLD=ON \
+      -DPGO_INSTRUMENT_LTO=Thin
+
+Then, to build the final optimized binary, build the stage2-clang++-bolt
+target:
+
+.. code-block:: console
+
+  $ ninja stage2-clang++-bolt
+
 3-Stage Non-Determinism
 =======================
 
 In the ancient lore of compilers non-determinism is like the multi-headed hydra.
 Whenever its head pops up, terror and chaos ensue.
 
 Historically one of the tests to verify that a compiler was deterministic would
 be a three stage build. The idea of a three stage build is you take your sources