diff options
| author | hunt <hunt> | 2006-03-27 19:37:25 +0000 |
|---|---|---|
| committer | hunt <hunt> | 2006-03-27 19:37:25 +0000 |
| commit | 57fe66f6d587630b757a1cd493f08763907f1f36 (patch) | |
| tree | 12e8b3d899b02f1fc5b19bef300d91592beedc78 | |
| parent | 7817aecdf972b7f3202d386ae97f2e77663261b9 (diff) | |
| download | systemtap-steved-57fe66f6d587630b757a1cd493f08763907f1f36.tar.gz systemtap-steved-57fe66f6d587630b757a1cd493f08763907f1f36.tar.xz systemtap-steved-57fe66f6d587630b757a1cd493f08763907f1f36.zip | |
*** empty log message ***
| -rw-r--r-- | runtime/ChangeLog | 15 |
1 files changed, 15 insertions, 0 deletions
diff --git a/runtime/ChangeLog b/runtime/ChangeLog index 0090a4f9..96ad8647 100644 --- a/runtime/ChangeLog +++ b/runtime/ChangeLog @@ -1,3 +1,18 @@ +2006-03-26 Martin Hunt <hunt@redhat.com> + + * bench2/bench.rb (Bench::run): Instead of loading the module + once and running against different numbers of threads, load + and unload the module each time. Then keep track of failures + per thread. Also use the new itest to keep the tests the same + size regardless of the number of threads. + (Bench::initialize): Make itest if needed. + (Bench::load, Stapbench::load): Increase buffer size to 8MB. + + * bench2/itest.c: Change arg to be the number of threads + instead of the number of millions of calls to make. + This makes it easy to divide the work among multiple threads + without making the test run many times longer. + 2006-03-09 Martin Hunt <hunt@redhat.com> * bench2: New directory containing a benchmark framework. |