No complaints from me, except that i've pretty much decided the first part of a new regression framework is a re-structuring of the config files so they can be used in the regression framework. In fact, if the config files were restructured, the new regression system would be pretty close.
Ali On Feb 7, 2012, at 8:45 PM, Steve Reinhardt wrote: > How about a new regression framework, assuming Ali isn't quietly putting > the finishing touches on what he started? > > On Tue, Feb 7, 2012 at 11:44 AM, Ali Saidi <[email protected]> wrote: > >> >> >> I think high on my list would be: >> >> 1. Making sampling via CPU >> switching work (e.g. switch between timing and detailed mode over and >> over again). This used to work, but some people have said it doesn't >> lately and a regression test to go with it. >> >> 2. Multi-threading at a >> system level (each System would run on it's own core and communicate >> across specially designed Etherlinks or whatever) >> >> 3. CPI stacks from >> the CPU model(s). >> >> Ali >> >> On 07.02.2012 12:12, Nilay Vaish wrote: >> >>> >> Can people comment on some of the features they would love to see in >>> >> gem5, that it lacks currently? >>> >>> -- >>> Nilay >>> >> _______________________________________________ >>> 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 >> > _______________________________________________ > 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
