From: Alexander S. <ast...@it...> - 2009-11-04 20:23:24
|
On Wed, Nov 4, 2009 at 2:42 PM, monipol <mo...@gm...> wrote: > Hello, all. What do you think of having a MaxCores parameter in > fink.conf and a %-field that could be explicitly used by maintainers > as in > > SetMAKEFLAGS: -j%{MaxCores} > > or something similar? In this way maintainers that have tested their > packages with -j could use that information without resorting to > `sysctl hw.ncpu` *and* users would be able to define their preferred > value in case they don't want all the cores to be used. Make sure that it's 'hw.activecpu' and not ncpu - it's possible to disable some cores, and only activecpu will change if you do that. > Also, Daniel Macks suggested having a MaxCores .info field, defaulting > to false, that could be set to true by packagers willing to use -j > with MaxCores. Quoting him: 'Seems like such a special-purpose %exp > that would only be used in one context, no need to pollute the whole > %exp table for every field parsing. For the corner-case where need to > pass -jX some way other than MAKEFLAGS env var, could just use the env > var in the *Script directly.' > > Bootstrap and fink configure could be modified to ask users which > value of MaxCores they prefer, showing the value of `sysctl hw.cpu` at > the prompt but not restricting whatever the user inputs to be less > than or equal to `sysctl hw.cpu`. Peter O'Gorman reports that -j3 is > faster than -j2 on his dual core machine. > > > Cheers, > > -- monipol > > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > fink-core mailing list > fin...@li... > http://news.gmane.org/gmane.os.apple.fink.core > |