I fully agree. I am not opposed to regressions staying if they add value. gzip just happened to be an easy target, and I'm open for suggestions.
I don't mind the CPU hours and MBs worth of stats changesets if it is for a good reason, but in this case I'm not sure it is. A better regression system would be fabulous. I'm eagerly waiting :-) Andreas On 27/05/2013 15:59, "Steve Reinhardt" <[email protected]> wrote: >I'm not necessarily opposed, it just seems somewhat arbitrary to single >out >gzip when the same logic could probably be applied to other tests. I >don't >run the long tests directly very much anyway, so the runtime doesn't >bother >me. > >Eventually we should have a more flexible regression system anyway... we >have a prototype internally at AMD but need to get some more experience >with it ourselves before releasing it. > >Steve > > >On Mon, May 27, 2013 at 7:48 AM, Andreas Hansson ><[email protected]>wrote: > >> It's one of the slower ones (1 - 1.5 CPU hours for each ISA, so x4). I >> have nothing against gzip in particular, I just do not see what it adds >> compared to e.g. bzip2. Do we believe that this test actually adds any >> value? >> >> The broader question would be: which of the tests actually add coverage >> and not just CPU hours? >> >> Another reason for pruning the regressions is the size of the stats >> updates. For some of the bus/memory related patches, each bump is 10MB+ >> (only stats.txt). >> >> Andreas >> >> >> >> On 27/05/2013 15:27, "Steve Reinhardt" <[email protected]> wrote: >> >> >Is there something that bothers you particularly about 00.gzip? It's >> >already in the 'long' set, correct? Aren't there some other equally >> >long-running tests? >> > >> >I don't have any particular attachment to gzip, but I don't want to >>see it >> >unfairly singled out... it seems no less valuable than many of our >>other >> >tests. >> > >> >Steve >> > >> > >> >On Mon, May 27, 2013 at 1:19 AM, Andreas Hansson >> ><[email protected]>wrote: >> > >> >> Hi all, >> >> >> >> Let me ask the same question in a different way: >> >> >> >> Could those that want to keep 00.gzip voice their opinion :-) >> >> >> >> I will create a patch to prune it in the meanwhile. >> >> >> >> Andreas >> >> >> >> >> >> On 22/05/2013 17:03, "Andreas Hansson" <[email protected]> >>wrote: >> >> >> >> >Hi everyone, >> >> > >> >> >How do people feel about pruning 00.gzip from the regressions? It >>takes >> >> >quite some time to run (1+ hour for each arch), and I struggle to >> >>justify >> >> >the time/energy spent. Does anyone think it adds value? >> >> > >> >> >Thanks, >> >> > >> >> >Andreas >> >> > >> >> >-- IMPORTANT NOTICE: The contents of this email and any attachments >>are >> >> >confidential and may also be privileged. If you are not the intended >> >> >recipient, please notify the sender immediately and do not disclose >>the >> >> >contents to any other person, use it for any purpose, or store or >>copy >> >> >the information in any medium. Thank you. >> >> >_______________________________________________ >> >> >gem5-dev mailing list >> >> >[email protected] >> >> >http://m5sim.org/mailman/listinfo/gem5-dev >> >> > >> >> >> >> >> >> -- IMPORTANT NOTICE: The contents of this email and any attachments >>are >> >> confidential and may also be privileged. If you are not the intended >> >> recipient, please notify the sender immediately and do not disclose >>the >> >> contents to any other person, use it for any purpose, or store or >>copy >> >>the >> >> information in any medium. Thank you. >> >> >> >> _______________________________________________ >> >> gem5-dev mailing list >> >> [email protected] >> >> http://m5sim.org/mailman/listinfo/gem5-dev >> >> >> >_______________________________________________ >> >gem5-dev mailing list >> >[email protected] >> >http://m5sim.org/mailman/listinfo/gem5-dev >> > >> >> >> -- IMPORTANT NOTICE: The contents of this email and any attachments are >> confidential and may also be privileged. If you are not the intended >> recipient, please notify the sender immediately and do not disclose the >> contents to any other person, use it for any purpose, or store or copy >>the >> information in any medium. Thank you. >> >> _______________________________________________ >> gem5-dev mailing list >> [email protected] >> http://m5sim.org/mailman/listinfo/gem5-dev >> >_______________________________________________ >gem5-dev mailing list >[email protected] >http://m5sim.org/mailman/listinfo/gem5-dev > -- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
