Cpu utilization

manohar · 12-16-2011, 12:26 AM

Hi,

Any one has idea how to find cpu utilization of functions in a process in linux platform.
top command will give process cpu utilization. I need to know cpu utilization of individual functions in the process..

jhwilliams · 12-16-2011, 01:32 AM

Try gprof or oprofile.

manohar · 12-16-2011, 04:49 AM

When i try to use "gprof" i am getting following message as output

Used :
gcc -lpthread -pg -g multicode.c -o gp-test
gprof gp-test

multicode.c has 3functions...

Quote:

Flat profile:

Each sample counts as 0.01 seconds.
no time accumulated

% cumulative self self total
time seconds seconds calls Ts/call Ts/call name

% the percentage of the total running time of the
time program used by this function.

cumulative a running sum of the number of seconds accounted
seconds for by this function and those listed above it.

self the number of seconds accounted for by this
seconds function alone. This is the major sort for this
listing.

calls the number of times this function was invoked, if
this function is profiled, else blank.

self the average number of milliseconds spent in this
ms/call function per call, if this function is profiled,
else blank.

total the average number of milliseconds spent in this
ms/call function and its descendents per call, if this
function is profiled, else blank.

name the name of the function. This is the minor sort
for this listing. The index shows the location of
the function in the gprof listing. If the index is
in parenthesis it shows where it would appear in
the gprof listing if it were to be printed.

dig2tech · 12-16-2011, 05:23 AM

Quote:

Originally Posted by manohar

Hi,

Any one has idea how to find cpu utilization of functions in a process in linux platform.
top command will give process cpu utilization. I need to know cpu utilization of individual functions in the process..

Hello Try with Sar or vmstat. It gives more info to you.

sundialsvcs · 12-16-2011, 07:31 AM

Probably the simplest way to do that sort of thing is to build instrumentation into the program itself: snapshot the timer before-and-after. Or, measure the system in terms of the work that it is actually doing, "for its customers" instead of "for the CPUs." One benefit of this strategy is that you can leave the instrumentation in, and you can glean statistics with minimal impact. The execution environment created by a profiler can be very different from the one that a program experiences in "normal" operation.

Another consideration is that, with multiple CPUs, many things can be happening at the same time such that a profiler's output is quite difficult to map to useful program improvements. I therefore prefer to try to understand a machine's operation in terms of what's going in and what's coming out (and when...), vs. counting the rotations of the internal gears.

Satyaveer Arya · 12-16-2011, 08:18 AM

Hello manohar,

Simply you can use mpstat or vmstat -a.

Output of mpstat on my RHEL5.3 Server:
07:46:36 PM CPU %user %nice %sys %iowait %irq %soft %steal %idle intr/s
07:46:36 PM all 0.77 0.00 0.12 0.68 0.00 0.05 0.00 98.38 694.26

Output of vmstat -a :
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
r b swpd free inact active si so bi bo in cs us sy id wa st
0 0 263512 46396 2105472 4364256 0 0 6 28 0 0 1 0 98 1 0