harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geir Magnusson Jr." <g...@pobox.com>
Subject Re: [build-test] regular runs of applications (was: Re: [performance] a few early benchmarks)
Date Tue, 28 Nov 2006 13:40:42 GMT

Mikhail Loenko wrote:
> 2006/11/28, Geir Magnusson Jr. <geir@pobox.com>:
>> Mikhail Loenko wrote:
>> > And we need to make sure that a patch that improves DaCapo by 0.1%
>> > does not lower down the SPEC by 10%
>> Absolutely :)
>> >
>> > We need to agree on a target function to optimize, like some 
>> weighted mean.
>> It would be useful to understand what SPECjbb is actually testing - can
>> someone describe it?
>> Then, like a financial portfolio, we can weight based on those factors...
> One more thing to work on is make sure that the workloads just run.
> It's not that easy to achieve some performance on the workloades if they
> just don't work :)
> I suggest that we setup regular runs of some workloads and report
> failures when regressions happen.
> Comments?

We've been proposing that for the CI system since the earth cooled. 
remember how useful running tomcat was to find the bugs related to 
thread manager signals?

The fact that we aren't there yet seems to be a lack of Round Tuits - 
we've been working to get the basic CI in place.  Once that's sound, I 
would imagine we can add scores of things like that if they are automated.


View raw message