On 11/11/2012 04:20 AM, Adam Borowski wrote: > Thomas Goirand says it's 400MB vs 700MB. I really doubt these numbers as > over 75% of all variables in all running programs would need to be pointers > so something must be amiss, but even far more conservative estimates you can > get by looking at executable size and/or some random structs in first piece > of code you have at hand, still give something quite nice to have.
My numbers are rough estimate of one's need for RAM in a VM which is mostly running Perl (Amavis + Spamassassin) and PHP. These number have nothing scientific, it's practical values. It demonstrate how much RAM is "wasted" (that can be debated during hours), but only to shows that in some cases, x32 would make a lot of sense. And it makes absolutely no doubts that x32 is valuable in this use case. > And, let's not argue about estimates Exactly! > , but let's have Daniel Schepler's work > uploaded SOMEWHERE that we can apt from, and take a look. Please yes! If it's the hosting space that is missing, I'll be happy to provide it before a better (eg: within Debian) place may be found (if interested, let me know). Thomas -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/509f4e96.1000...@debian.org