This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] New alias grouping heuristic
On Wed, 2004-02-25 at 09:52, Vladimir Makarov wrote:
> The second, I found that SPEC results are pretty stable if you run it
> right on a free machine. I have stable results in range 2-4 for score
> about 800 for P4. Although the volatility could be more for single
> tests. But the geometry meaning is pretty the same.
>
Perhaps, but Mike was quoting individual tests. Thinks like mcf are
*really* volatile. At least on the P3 and P4 boxes I run it every day.
Look at the daily graphs in
http://people.redhat.com/dnovillo/spec2000/tree-ssa-branch/individual-run-ratio.html
The only tests that seem somewhat stable day-to-day are gap, crafty and
most of the FP tests. Granted, the daily runs use only one run and the
machine is not isolated to the point of being in single user mode. What
the scripts do is stop cron, at, sendmail, disallow logins, SIGSTOP
non-root processes and build everything in a local filesystem.
Diego.