aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorDaniel Eklöf <daniel@ekloef.se>2020-12-21 12:01:11 +0000
committerDaniel Eklöf <daniel@ekloef.se>2020-12-21 12:01:11 +0000
commita83f2448bcf420365df401674d42ca1ac1c692e2 (patch)
treeb60305dee869a18331ca287a0077d975bd45635a
parentchangelog: use -std=c11 instead of -std=c18 (diff)
parentinstall.md: mention that packagers should not use -march=native (diff)
Merge branch 'clang-wno-profile-instr-unprofile'
-rw-r--r--INSTALL.md9
1 files changed, 8 insertions, 1 deletions
diff --git a/INSTALL.md b/INSTALL.md
index f27cb475..f5cbc184 100644
--- a/INSTALL.md
+++ b/INSTALL.md
@@ -142,6 +142,10 @@ optimization.
In addition to being faster, PGO builds also tend to be smaller than
regular `-O3` builds.
+**Note**: if you are packaging foot, you should probably remove
+`-march=native`, as that may cause gcc/clang to emit instructions that
+may not be available on other computers.
+
#### Size optimized
@@ -186,7 +190,10 @@ later have a regression where PGO with `-O2` is **much** slower.
If you are using Clang instead of GCC, use the following `CFLAGS` instead:
```sh
-export CFLAGS="$CFLAGS -O3 -march=native -Wno-ignored-optimization-argument -Wno-profile-instr-out-of-date"
+export CFLAGS="$CFLAGS -O3 -march=native \
+ -Wno-ignored-optimization-argument \
+ -Wno-profile-instr-out-of-date \
+ -Wno-profile-instr-unprofiled"
```
Then, tell meson we want to _generate_ profiling data, and build: