Philip Mucci wrote: > Stefane, > > Can we roll this into the libpfm distro? I've been writing tests (as you > know), but it would be great to have a framework for this. > > Phil > > On Apr 22, 2008, at 4:51 PM, William Cohen wrote: > >> stephane eranian wrote: >>> Robert, >>> On Tue, Apr 22, 2008 at 11:54 AM, Robert Richter >>> <[EMAIL PROTECTED]> wrote: >>>> Stephane, >>>> >>>> >>>> On 21.04.08 19:51:21, stephane eranian wrote: >>>> > Someone else also altered me on these problems. I think I fixed >>>> all of them >>>> > no, incl. install_examples. >>>> > Could you try again? >>>> >>>> todays Git (Linux 2.6.25) and CVS is working fine (x86). >>>> >>> Good. >>> I will release the 2.6.25 patch later this week, so please everyone >>> stress test the code has much as you can until then. >>> Thanks. >> >> Hi Stephane, >> >> I manual ran the examples in the libpfm/examples with the 2.6.25 >> kernel and a current cvs checkout of libpfm. Nothing crashed. :) >> >> I have attached a dejagnu testsuite that does some very simple smoke >> tests. It is run with: >> >> runtest --tool pfmon >> >> It checks that event counting is repeatable and that multiple threads >> do not interfere with each other. >> >> -Will >> <testsuite-0.1-20080422.tar.gz>------------------------------------------------------------------------- >> >> >> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference >> Don't miss this year's exciting event. There's still time to save $100. >> Use priority code J8TL2D2. >> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone_______________________________________________ >> >> >> perfmon2-devel mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/perfmon2-devel >
Hi Phil, I am all for rolling some automated tests into libpfm. It would good to have some easy way (e.g. "make check") for people to check that things appear to be functioning as expected. SystemTap has a systemta-testsuite rpm built that has all the test related stuff in a simple installable form, so people that install the regular systemtap rpm can test things on their local environment. Given that the performance monitoring hardware is very implementation dependent, making it easy for people to identify that something in libpfm isn't working on their particular hardware would greatly help to make sure that code is really being exercised. -Will ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ perfmon2-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
