Greenplum或PostgreSQL的编译配置文件中有一个选项是 --enable-profiling，这个选项会打开GCC的-pg参数。产生可以被gprof用于分析进程的代码。
如下：
less configure
# enable profiling if --enable-profiling
if test "$enable_profiling" = yes && test "$ac_cv_prog_cc_g" = yes; then
if test "$GCC" = yes; then
$as_echo "#define PROFILE_PID_DIR 1" >>confdefs.h
CFLAGS="$CFLAGS -pg $PLATFORM_PROFILE_FLAGS"
else
as_fn_error $? "--enable-profiling is supported only when using GCC" "$LINENO" 5
fi
fi
less src/template/linux
# If --enable-profiling is specified, we need -DLINUX_PROFILE
PLATFORM_PROFILE_FLAGS="-DLINUX_PROFILE"
man gcc
-pg Generate extra code to write profile information suitable for the analysis program gprof. You must use this option when compiling the source files you want data about, and you must also use it when linking.
诊断代码被放在$PGDATA/gprof目录中，每个进程都会产生一个子目录，里面会有一个gmon.out文件。
查看这个文件的信息：
#gprof -b /home/digoal/gpdb/bin/postgres gmon.out
Flat profile:
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls ms/call ms/call name
80.00 0.04 0.04 91 0.44 0.44 FileRepVerify_ComputeFileHash
20.00 0.05 0.01 1 10.00 10.00 ChangeTracking_WriteBuffer
0.00 0.05 0.00 750562 0.00 0.00 ChangeTracking_GetRelationChangeInfoFromXlog
0.00 0.05 0.00 670756 0.00 0.00 cdbpullup_colIdx
0.00 0.05 0.00 253200 0.00 0.00 cdbpullup_isExprCoveredByTargetlist
0.00 0.05 0.00 239413 0.00 0.00 cdbpullup_targetlist
......
有点类似oprofile的输出。
gprof的用法可以参考man gprof。