Comparing 1.4.0fcs with promoted build b08, running c2 gives us a 1% regression for Scimark on solaris sparc, a 3% regression for SpecJBB on windows, and a 2% regression for SpecJBB on linux. All regressions are significant and have a p value of 0.0001

You can see the log files at:
1.4.0
Solaris Sparc
/java/alacrity/a1.02/prod/alacrity/logs/1.4.0/p/fcs/2002-01-30/000/solaris/sparc-US-II/server/*
Windows
/java/alacrity/a1.02/prod/alacrity/logs/1.4.0/p/fcs/2002-02-07/000/windows/i386/server/*
Linux
/java/alacrity/a1.02/prod/alacrity/logs/1.4.0/p/fcs/2002-01-30/000/linux/i386/server/*

and
1.4.1 b08
Solaris Sparc
/java/alacrity/a1.02/prod/alacrity/logs/1.4.1/p/b08/2002-04-09/000/solaris/sparc-US-II/server/*
Windows
/java/alacrity/a1.02/prod/alacrity/logs/1.4.1/p/b08/2002-04-09/000/windows/i386/server/*
Linux
/java/alacrity/a1.02/prod/alacrity/logs/1.4.1/p/b08/2002-04-09/000/linux/i386/server/*

Just to update on this bug:
Now b09 shows a regression in linux scimark at about 2%. It was 1% regression on b08 but it was not significant so a bug was not filed. Now the regression has increased and has become statiscally significant.
You can find the b09 logs at:
/java/alacrity/latest/prod/alacrity/logs/1.4.1/p/b09/2002-04-16/000/linux/i386/server/*/results.scimark


###@###.### 2002-04-17