Measuring Execution Time
|
View:
New views
5 Messages
—
Rating Filter:
Alert me
|
 |
Measuring Execution Time
by Peter Vandrish
::
Rate this Message:
Hi,
I am trying to measure the execution time of my program using the following code: #include <ctime> ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ gumstix-users mailing list gumstix-users@... https://lists.sourceforge.net/lists/listinfo/gumstix-users |
|
|
Re: Measuring Execution Time
by Dave Hylands
::
Rate this Message:
Hi Pete,
> I am trying to measure the execution time of my program using the following > code: > #include <ctime> > clock_t start,finish; > double time; > > > start = clock(); > sort something > finish = clock(); > > > time = (double(finish)-double(start))/CLOCKS_PER_SEC; > > It does not seem accurate at all. It reports that the Gumstix > CLOCKS_PER_SECOND is 1000000. > > It also tells me the double(finish) is 30000 and the double(start) is 20000, > it doesnt seem likely > > > > that my program executes in a nice round number of 10000 CLOCK TICKS. > > Does anyone know hoe I can do this well. Try using the gettimeofday function. <http://linux.die.net/man/2/gettimeofday> On some version of linux (running on the PXA processors) this will return sub-microsecond resolution, while on others it will return 10 millisecond resolution. Note that gettimeofday will return wall time, as opposed to processor time. You may also want to look at the times function: <http://linux.die.net/man/2/times> -- Dave Hylands Vancouver, BC, Canada http://www.DaveHylands.com/ ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ gumstix-users mailing list gumstix-users@... https://lists.sourceforge.net/lists/listinfo/gumstix-users |
|
|
Re: Measuring Execution Time
by Peter Vandrish
::
Rate this Message:
Thanks again Dave!
On Thu, May 22, 2008 at 2:00 PM, Dave Hylands <dhylands@...> wrote: Hi Pete, ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ gumstix-users mailing list gumstix-users@... https://lists.sourceforge.net/lists/listinfo/gumstix-users |
|
|
Re: Measuring Execution Time
by adrian cockcroft
::
Rate this Message:
This data does make sense.
CPU time is charged by the clock interrupt at 100Hz in 10ms increments. Your program had used 2 ticks at start and 3 ticks at finish. The time is being scaled to microseconds. So your sort took between 1 and 19ms of CPU time. If you run the code lots of times or loop the test and average the result you can get a more accurate measure. Adrian On Thu, May 22, 2008 at 8:15 AM, Peter Vandrish <peter.vandrish@...> wrote: > Hi, > I am trying to measure the execution time of my program using the following > code: > > #include <ctime> > clock_t start,finish; > double time; > > start = clock(); > sort something > finish = clock(); > > > time = (double(finish)-double(start))/CLOCKS_PER_SEC; > > It does not seem accurate at all. It reports that the Gumstix > CLOCKS_PER_SECOND is 1000000. > > It also tells me the double(finish) is 30000 and the double(start) is 20000, > it doesnt seem likely > > that my program executes in a nice round number of 10000 CLOCK TICKS. > > Does anyone know hoe I can do this well. > > Thanks, > > Pete > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > gumstix-users mailing list > gumstix-users@... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ gumstix-users mailing list gumstix-users@... https://lists.sourceforge.net/lists/listinfo/gumstix-users |
|
|
Re: Measuring Execution Time
by peteb
::
Rate this Message:
If you just want to measure the time your program was executing, just
run it with the time command at the command prompt. That returns the
clock time and CPU time used. Very handy. Won't work for measuring bits
of the app, but otherwise might be useful as it doesn't require you to
add time code to your application.
# time app # # real Xm XXs # user Xm XXs # sys Xm XXs Peter Vandrish wrote: Thanks again Dave! -- Pete Funny=Truth Truth=Beauty So take it as a complement when people say you look Funny. ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ gumstix-users mailing list gumstix-users@... https://lists.sourceforge.net/lists/listinfo/gumstix-users |
| Free Forum Powered by Nabble | Forum Help |