That's an excellent idea and yes, that seems like a good course of action.
Ali On 23.01.2013 14:04, nathan binkert wrote: > With the new system, for this case, dumping a subset of stats would > not be all that difficult. If you're dumping that frequently, it > doesn't seem to be a huge burden to expect people to narrow down the > subset of stats that they're interested in if they want speed. > > On Wed, Jan 23, 2013 at 11:59 AM, Ali Saidi <[email protected]> wrote: > >> Yes. Pretty much the question in my mind is how many people are dumping stats at a rate greater than once every 100k instructions? Thanks, Ali On 23.01.2013 11:53, Steve Reinhardt wrote: >> >>> Right, >> I'd say the main concern here is for people whose usage modelincorporates fairly frequent periodic stats dumps (who are probably more >> >>> likely to want the SQL output anyway). Steve On Wed, >> Jan 23, 2013 at 9:27 AM, nathan binkert <[email protected]> wrote:It's worth making the point that this is only when dumping the stats. Stats collection hasn't changed. If performance of the dump is really a concern, I'm sure I can offer some suggestions to improve things. On Wed, Jan 23, 2013 at 9:22 AM, Andreas Hansson <[email protected]> wrote: > > _______________________________________________ > gem5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/gem5-dev [1] Links: ------ [1] http://m5sim.org/mailman/listinfo/gem5-dev _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
