On 11/02/2016 05:30 PM, Ed Greshko wrote: > > > On 11/03/16 07:10, jd1008 wrote: >> >> >> On 11/02/2016 11:59 AM, Rick Stevens wrote: >>> On 11/02/2016 10:41 AM, jd1008 wrote: >>>> Fedora-Workstation-Live-x86_64-24-1.2.iso >>>> <http://mirrors.kernel.org/fedora/releases/24/Workstation/x86_64/iso/Fedora-Workstation-Live-x86_64-24-1.2.iso> >>>> >>>> 14-Jun-2016 17:13 1G >>>> Fedora-Workstation-Live-i386-24-1.2.iso >>>> <http://mirrors.kernel.org/fedora/releases/24/Workstation/i386/iso/Fedora-Workstation-Live-i386-24-1.2.iso> >>>> >>>> 14-Jun-2016 17:27 2G >>>> >>>> Why? >>> Have you looked at the file lists? I haven't (I don't have any 32-bit >>> machines), but a first guess may be that the 32-bit version has both >>> PAE and non-PAE kernels, libraries, etc., so you have more "stuff" to >>> support the hardware environment adequately. >> >> On the face of it, it sounds logical, but it makes no sense. >> I can understand that 64bit systems are supposed to support 32bit apps, >> but I have never come across a 32bit system that supports 64bit apps. >> Apps do not depend on low level PAE OS architecture. > > It makes perfect sense. > > The PAE kernel allows for access of up to 64GB of RAM from a max of 4GB on 32 bit systems > whose CPU's offer the PAE extension. Nothing to do with support of 32-bit v.s. 64-bit apps. > > If you mount the WS-32bit iso and then mount the squashfs under LiveOS and then mount the > rootfs.img under LiveOS of the mounted squashfs you'd see.... > > [egreshko@meimei iso2]$ ls > bin dev home lost+found mnt proc run srv tmp var > boot etc lib media opt root sbin sys usr > [egreshko@meimei iso2]$ ls boot > config-4.5.5-300.fc24.i686 initramfs-4.5.5-300.fc24.i686+PAE.img > config-4.5.5-300.fc24.i686+PAE initrd-plymouth.img > efi memtest86+-5.01 > elf-memtest86+-5.01 System.map-4.5.5-300.fc24.i686 > extlinux System.map-4.5.5-300.fc24.i686+PAE > grub2 vmlinuz-4.5.5-300.fc24.i686 > initramfs-4.5.5-300.fc24.i686.img vmlinuz-4.5.5-300.fc24.i686+PAE > > While with the WS-64bit iso you'd see > > [egreshko@meimei iso2]$ ls > bin dev home lib64 media opt root sbin sys usr > boot etc lib lost+found mnt proc run srv tmp var > [egreshko@meimei iso2]$ ls boot > config-4.5.5-300.fc24.x86_64 initramfs-4.5.5-300.fc24.x86_64.img > efi initrd-plymouth.img > elf-memtest86+-5.01 memtest86+-5.01 > extlinux System.map-4.5.5-300.fc24.x86_64 > grub2 vmlinuz-4.5.5-300.fc24.x86_64 Thanks for elucidating that for me, Ed. I believe there's also differences in the library trees, too, for PAE and non-PAE. JD, what Ed's saying is that 32-bit CPUs that do NOT have the PAE extensions can only access 4GB of RAM. The PAE extensions present on some CPUs permit access to 64GB of RAM. There are kernel differences between the two and I think (although I'm not certain) that there may be two different sets of libraries as well (or at least code in them to handle the differences). That can easily explain the size differences. The only alternative would be for Fedora to offer two different sets of 32-bit disks, non-PAE and PAE-enabled. There's enough confusion with the disk sets as it stands now. I can sure see why they don't do it. ---------------------------------------------------------------------- - Rick Stevens, Systems Engineer, AllDigital ricks@xxxxxxxxxxxxxx - - AIM/Skype: therps2 ICQ: 226437340 Yahoo: origrps2 - - - - Let us think the unthinkable. Let us do the undoable. Let us - - prepare to grapple with the ineffable itself, and see if we may - - not eff it up after all. - - -- Douglas Adams - ---------------------------------------------------------------------- _______________________________________________ users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
