xref: /freebsd/share/doc/papers/kerntune/2.t (revision afe61c15161c324a7af299a9b8457aba5afc92db)
Copyright (c) 1984 M. K. McKusick
Copyright (c) 1984 The Regents of the University of California.
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
3. All advertising materials mentioning features or use of this software
must display the following acknowledgement:
This product includes software developed by the University of
California, Berkeley and its contributors.
4. Neither the name of the University nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
SUCH DAMAGE.

@(#)2.t 1.3 (Berkeley) 11/8/90

The gprof Profiler

The purpose of the gprof profiling tool is to help the user evaluate alternative implementations of abstractions. The gprof design takes advantage of the fact that the kernel though large, is structured and hierarchical. We provide a profile in which the execution time for a set of routines that implement an abstraction is collected and charged to that abstraction. The profile can be used to compare and assess the costs of various implementations [Graham82] [Graham83]. Data presentation

The data is presented to the user in two different formats. The first presentation simply lists the routines without regard to the amount of time their descendants use. The second presentation incorporates the call graph of the kernel. The Flat Profile

The flat profile consists of a list of all the routines that are called during execution of the kernel, with the count of the number of times they are called and the number of seconds of execution time for which they are themselves accountable. The routines are listed in decreasing order of execution time. A list of the routines that are never called during execution of the kernel is also available to verify that nothing important is omitted by this profiling run. The flat profile gives a quick overview of the routines that are used, and shows the routines that are themselves responsible for large fractions of the execution time. In practice, this profile usually shows that no single function is overwhelmingly responsible for the total time of the kernel. Notice that for this profile, the individual times sum to the total execution time. The Call Graph Profile

Ideally, we would like to print the call graph of the kernel, but we are limited by the two-dimensional nature of our output devices. We cannot assume that a call graph is planar, and even if it is, that we can print a planar version of it. Instead, we choose to list each routine, together with information about the routines that are its direct parents and children. This listing presents a window into the call graph. Based on our experience, both parent information and child information is important, and should be available without searching through the output. Figure 1 shows a sample gprof entry. .KF L

called/total parents
index %time self descendants called+self name index
called/total children
0.20 1.20 4/10 \s-1CALLER1\s+1 [7]
0.30 1.80 6/10 \s-1CALLER2\s+1 [1]
[2] 41.5 0.50 3.00 10+4 \s-1EXAMPLE\s+1 [2]
1.50 1.00 20/40 \s-1SUB1\s+1 <cycle1> [4]
0.00 0.50 1/5 \s-1SUB2\s+1 [9]
0.00 0.00 0/5 \s-1SUB3\s+1 [11]

Figure 1. Profile entry for \s-1EXAMPLE\s+1. .KE

The major entries of the call graph profile are the entries from the flat profile, augmented by the time propagated to each routine from its descendants. This profile is sorted by the sum of the time for the routine itself plus the time inherited from its descendants. The profile shows which of the higher level routines spend large portions of the total execution time in the routines that they call. For each routine, we show the amount of time passed by each child to the routine, which includes time for the child itself and for the descendants of the child (and thus the descendants of the routine). We also show the percentage these times represent of the total time accounted to the child. Similarly, the parents of each routine are listed, along with time, and percentage of total routine time, propagated to each one.

Cycles are handled as single entities. The cycle as a whole is shown as though it were a single routine, except that members of the cycle are listed in place of the children. Although the number of calls of each member from within the cycle are shown, they do not affect time propagation. When a child is a member of a cycle, the time shown is the appropriate fraction of the time for the whole cycle. Self-recursive routines have their calls broken down into calls from the outside and self-recursive calls. Only the outside calls affect the propagation of time.

The example shown in Figure 2 is the fragment of a call graph corresponding to the entry in the call graph profile listing shown in Figure 1. .KF L .so fig2.pic

Figure 2. Example call graph fragment. .KE

The entry is for routine \s-1EXAMPLE\s+1, which has the Caller routines as its parents, and the Sub routines as its children. The reader should keep in mind that all information is given with respect to \s-1EXAMPLE\s+1. The index in the first column shows that \s-1EXAMPLE\s+1 is the second entry in the profile listing. The \s-1EXAMPLE\s+1 routine is called ten times, four times by \s-1CALLER1\s+1, and six times by \s-1CALLER2\s+1. Consequently 40% of \s-1EXAMPLE\s+1's time is propagated to \s-1CALLER1\s+1, and 60% of \s-1EXAMPLE\s+1's time is propagated to \s-1CALLER2\s+1. The self and descendant fields of the parents show the amount of self and descendant time \s-1EXAMPLE\s+1 propagates to them (but not the time used by the parents directly). Note that \s-1EXAMPLE\s+1 calls itself recursively four times. The routine \s-1EXAMPLE\s+1 calls routine \s-1SUB1\s+1 twenty times, \s-1SUB2\s+1 once, and never calls \s-1SUB3\s+1. Since \s-1SUB2\s+1 is called a total of five times, 20% of its self and descendant time is propagated to \s-1EXAMPLE\s+1's descendant time field. Because \s-1SUB1\s+1 is a member of cycle 1, the self and descendant times and call count fraction are those for the cycle as a whole. Since cycle 1 is called a total of forty times (not counting calls among members of the cycle), it propagates 50% of the cycle's self and descendant time to \s-1EXAMPLE\s+1's descendant time field. Finally each name is followed by an index that shows where on the listing to find the entry for that routine. Profiling the Kernel

It is simple to build a 4.2BSD kernel that will automatically collect profiling information as it operates simply by specifying the -p option to config\|(8) when configuring a kernel. The program counter sampling can be driven by the system clock, or by an alternate real time clock. The latter is highly recommended as use of the system clock results in statistical anomalies in accounting for the time spent in the kernel clock routine.

Once a profiling system has been booted statistic gathering is handled by kgmon\|(8). Kgmon allows profiling to be started and stopped and the internal state of the profiling buffers to be dumped. Kgmon can also be used to reset the state of the internal buffers to allow multiple experiments to be run without rebooting the machine. The profiling data can then be processed with gprof\|(1) to obtain information regarding the system's operation.

A profiled system is about 5-10% larger in its text space because of the calls to count the subroutine invocations. When the system executes, the profiling data is stored in a buffer that is 1.2 times the size of the text space. All the information is summarized in memory, it is not necessary to have a trace file being continuously dumped to disk. The overhead for running a profiled system varies; under normal load we see anywhere from 5-25% of the system time spent in the profiling code. Thus the system is noticeably slower than an unprofiled system, yet is not so bad that it cannot be used in a production environment. This is important since it allows us to gather data in a real environment rather than trying to devise synthetic work loads.