• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1=============================
2Advanced Build Configurations
3=============================
4
5.. contents::
6   :local:
7
8Introduction
9============
10
11`CMake <http://www.cmake.org/>`_ is a cross-platform build-generator tool. CMake
12does not build the project, it generates the files needed by your build tool
13(GNU make, Visual Studio, etc.) for building LLVM.
14
15If **you are a new contributor**, please start with the :doc:`GettingStarted` or
16:doc:`CMake` pages. This page is intended for users doing more complex builds.
17
18Many of the examples below are written assuming specific CMake Generators.
19Unless otherwise explicitly called out these commands should work with any CMake
20generator.
21
22Bootstrap Builds
23================
24
25The Clang CMake build system supports bootstrap (aka multi-stage) builds. At a
26high level a multi-stage build is a chain of builds that pass data from one
27stage into the next. The most common and simple version of this is a traditional
28bootstrap build.
29
30In a simple two-stage bootstrap build, we build clang using the system compiler,
31then use that just-built clang to build clang again. In CMake this simplest form
32of a bootstrap build can be configured with a single option,
33CLANG_ENABLE_BOOTSTRAP.
34
35.. code-block:: console
36
37  $ cmake -G Ninja -DCLANG_ENABLE_BOOTSTRAP=On <path to source>
38  $ ninja stage2
39
40This command itself isn't terribly useful because it assumes default
41configurations for each stage. The next series of examples utilize CMake cache
42scripts to provide more complex options.
43
44The clang build system refers to builds as stages. A stage1 build is a standard
45build using the compiler installed on the host, and a stage2 build is built
46using the stage1 compiler. This nomenclature holds up to more stages too. In
47general a stage*n* build is built using the output from stage*n-1*.
48
49Apple Clang Builds (A More Complex Bootstrap)
50=============================================
51
52Apple's Clang builds are a slightly more complicated example of the simple
53bootstrapping scenario. Apple Clang is built using a 2-stage build.
54
55The stage1 compiler is a host-only compiler with some options set. The stage1
56compiler is a balance of optimization vs build time because it is a throwaway.
57The stage2 compiler is the fully optimized compiler intended to ship to users.
58
59Setting up these compilers requires a lot of options. To simplify the
60configuration the Apple Clang build settings are contained in CMake Cache files.
61You can build an Apple Clang compiler using the following commands:
62
63.. code-block:: console
64
65  $ cmake -G Ninja -C <path to clang>/cmake/caches/Apple-stage1.cmake <path to source>
66  $ ninja stage2-distribution
67
68This CMake invocation configures the stage1 host compiler, and sets
69CLANG_BOOTSTRAP_CMAKE_ARGS to pass the Apple-stage2.cmake cache script to the
70stage2 configuration step.
71
72When you build the stage2-distribution target it builds the minimal stage1
73compiler and required tools, then configures and builds the stage2 compiler
74based on the settings in Apple-stage2.cmake.
75
76This pattern of using cache scripts to set complex settings, and specifically to
77make later stage builds include cache scripts is common in our more advanced
78build configurations.
79
80Multi-stage PGO
81===============
82
83Profile-Guided Optimizations (PGO) is a really great way to optimize the code
84clang generates. Our multi-stage PGO builds are a workflow for generating PGO
85profiles that can be used to optimize clang.
86
87At a high level, the way PGO works is that you build an instrumented compiler,
88then you run the instrumented compiler against sample source files. While the
89instrumented compiler runs it will output a bunch of files containing
90performance counters (.profraw files). After generating all the profraw files
91you use llvm-profdata to merge the files into a single profdata file that you
92can feed into the LLVM_PROFDATA_FILE option.
93
94Our PGO.cmake cache script automates that whole process. You can use it by
95running:
96
97.. code-block:: console
98
99  $ cmake -G Ninja -C <path_to_clang>/cmake/caches/PGO.cmake <source dir>
100  $ ninja stage2-instrumented-generate-profdata
101
102If you let that run for a few hours or so, it will place a profdata file in your
103build directory. This takes a really long time because it builds clang twice,
104and you *must* have compiler-rt in your build tree.
105
106This process uses any source files under the perf-training directory as training
107data as long as the source files are marked up with LIT-style RUN lines.
108
109After it finishes you can use “find . -name clang.profdata” to find it, but it
110should be at a path something like:
111
112.. code-block:: console
113
114  <build dir>/tools/clang/stage2-instrumented-bins/utils/perf-training/clang.profdata
115
116You can feed that file into the LLVM_PROFDATA_FILE option when you build your
117optimized compiler.
118
119The PGO came cache has a slightly different stage naming scheme than other
120multi-stage builds. It generates three stages; stage1, stage2-instrumented, and
121stage2. Both of the stage2 builds are built using the stage1 compiler.
122
123The PGO came cache generates the following additional targets:
124
125**stage2-instrumented**
126  Builds a stage1 x86 compiler, runtime, and required tools (llvm-config,
127  llvm-profdata) then uses that compiler to build an instrumented stage2 compiler.
128
129**stage2-instrumented-generate-profdata**
130  Depends on "stage2-instrumented" and will use the instrumented compiler to
131  generate profdata based on the training files in <clang>/utils/perf-training
132
133**stage2**
134  Depends of "stage2-instrumented-generate-profdata" and will use the stage1
135  compiler with the stage2 profdata to build a PGO-optimized compiler.
136
137**stage2-check-llvm**
138  Depends on stage2 and runs check-llvm using the stage2 compiler.
139
140**stage2-check-clang**
141  Depends on stage2 and runs check-clang using the stage2 compiler.
142
143**stage2-check-all**
144  Depends on stage2 and runs check-all using the stage2 compiler.
145
146**stage2-test-suite**
147  Depends on stage2 and runs the test-suite using the stage3 compiler (requires
148  in-tree test-suite).
149
1503-Stage Non-Determinism
151=======================
152
153In the ancient lore of compilers non-determinism is like the multi-headed hydra.
154Whenever it's head pops up, terror and chaos ensue.
155
156Historically one of the tests to verify that a compiler was deterministic would
157be a three stage build. The idea of a three stage build is you take your sources
158and build a compiler (stage1), then use that compiler to rebuild the sources
159(stage2), then you use that compiler to rebuild the sources a third time
160(stage3) with an identical configuration to the stage2 build. At the end of
161this, you have a stage2 and stage3 compiler that should be bit-for-bit
162identical.
163
164You can perform one of these 3-stage builds with LLVM & clang using the
165following commands:
166
167.. code-block:: console
168
169  $ cmake -G Ninja -C <path_to_clang>/cmake/caches/3-stage.cmake <source dir>
170  $ ninja stage3
171
172After the build you can compare the stage2 & stage3 compilers. We have a bot
173setup `here <http://lab.llvm.org:8011/builders/clang-3stage-ubuntu>`_ that runs
174this build and compare configuration.
175