m...@linux.it writes: > On Mar 16, Simon Richter <s...@debian.org> wrote: > >> Well, it would get i386/amd64 in line with sparc/sparc64, powerpc/powerpc64 >> and s390/s390x. That would allow us to get rid of a lot of specianl cases, >> including the hack for libc6-386. > I think it would be very helpful if somebody could summarize why a > multiarch system is useful, except for the obvious case of installing > proprietary i386 software on amd64 systems.
People who use those systems frequently could probably provide more cases, but I know of three cases beyond running "qualified" or "supported" 32-bit binaries on 64-bit systems): - On platforms where the 32-bit mode is not so register-poor as i386, using 32-bit binaries can be significantly faster and less cache-disruptive than 64-bit binaries because they need less memory bandwidth. Even on amd64 systems, some executables have this trait. - Some users have code that does not function properly on systems with sizeof(int) != sizeof(void*). Multiarch gives them a chance to migrate this code at their own pace. - For environments with certain configuration management or regulatory requirements, qualifying new binaries is a considerable cost -- having nothing to do with whether the software is "proprietary" -- and multiarch makes it easier for them also to migrate at a pace that makes sense to them. Michael Poole -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
