From: phung t. <nh_...@ya...> - 2003-06-21 16:18:12
|
Hi all, I am trying to measure the performance when running original jikes. And then I try to measure the performance when applying a new optimizing technique. However, the results are not stable in both versions. The following is from the original version ======= _228_jack Finished in 105.223 secs ======= _228_jack Finished in 45.667 secs ======= _228_jack Finished in 49.946 secs ======= _228_jack Finished in 46.671 secs ======= _228_jack Finished in 63.951 secs Is this result normal? The result was achieved from 1 single RVM run in o1 level. I understand the first run slow because it includes compilation time but the others are not stable either. is there anyone getting the same problem? how could you solve it? Or simply choosing the best run? I highly appreciate any help Thanks Phung __________________________________ Do you Yahoo!? SBC Yahoo! DSL - Now only $29.95 per month! http://sbc.yahoo.com |