CMake: enable LTO in release mode

Link-Time Optimization (LTO) is a mechanism that enables the compiler to
optimize across translation unit boundaries. In particular, it enables
cross-file function inlining. This is present and mature in the majority of
contemporary compilers and switching it on has few downsides.

Some performance numbers:

  ┌───────┬──────────────────┬──────────────────┬──────┐
  │       │ before           │ after            │ diff │
  ╞═══════╪══════════════════╪══════════════════╪══════╡
  │ 1652¹ │ 17.88s           │ 17.85s           │  -0% │
  │       │ 50.4MB peak RSS  │ 50.5MB peak RSS  │  +0% │
  ├───────┼──────────────────┼──────────────────┼──────┤
  │ 1718² │ 2m21s            │ 2m13s            │  -6% │
  │       │ 15.8MB peak RSS  │ 15.9 MB peak RSS │  +1% │
  ├───────┼──────────────────┼──────────────────┼──────┤
  │ 1864³ │ 13.26s           │ 13.07s           │  -1% │
  │       │ 462.0MB peak RSS │ 461.8MB peak RSS │  -0% │
  ├───────┼──────────────────┼──────────────────┼──────┤
  │ 2064⁴ │ 11m42s           │ 11m30s           │  -2% │
  │       │ 1.26GB peak RSS  │ 1.26GB peak RSS  │  -0% │
  ├───────┼──────────────────┼──────────────────┼──────┤
  │ 2095⁵ │ 2m18s            │ 2m19s            │  +1% │
  │       │ 92.3MB peak RSS  │ 92.1MB peak RSS  │  -0% │
  └───────┴──────────────────┴──────────────────┴──────┘

¹ The test case from https://gitlab.com/graphviz/graphviz/-/issues/1652 run as
  `neato -Tsvg -o /dev/null 1652.dot`.
² swedish-flat.dot Magnus attached to
  https://gitlab.com/graphviz/graphviz/-/issues/1718 run as
  `circo -Tsvg -o /dev/null swedish-flag.dot`.
³ The test case from https://gitlab.com/graphviz/graphviz/-/issues/1864 run as
  `twopi -Tsvg -o /dev/null 1864.dot`.
⁴ The test case from https://gitlab.com/graphviz/graphviz/-/issues/2064 run as
  `dot -Gnslimit=2 -Gnslimit1=2 -Gmaxiter=5000 -Tsvg -o /dev/null 2064.dot`.
⁵ The tests/2095.dot test case from prior to minimization
  (3819821ea70fae730dd224936628ed3929b03531). Run as
  `dot -Tsvg -o /dev/null 2095.dot`.

diff --git a/CMakeLists.txt b/CMakeLists.txt
index fc3a377b2db99571d3a2a2721f4ffa03447f6135..e425ae38c055882d26e45e05f32260f82d837980 100644 (file)
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -220,6 +220,16 @@ set(CMAKE_CXX_STANDARD_REQUIRED ON)
 
 set(CMAKE_POSITION_INDEPENDENT_CODE ON)
 
+# enable LTO in release builds
+include(CheckIPOSupported)
+check_ipo_supported(RESULT HAVE_IPO)
+if(CMAKE_BUILD_TYPE STREQUAL "Release" OR
+   CMAKE_BUILD_TYPE STREQUAL "RelWithDebInfo")
+  if(HAVE_IPO)
+    set(CMAKE_INTERPROCEDURAL_OPTIMIZATION ON)
+  endif()
+endif()
+
 if(${CMAKE_CXX_COMPILER_ID} STREQUAL MSVC)
   # MSVC warning C4996 mostly fires on completely valid code. The changes
   # proposed in the warning text often seriously compromise the code