[Leaplist] FC6 conversion from Ubuntu?

Dan Cherry dscherry at bellsouth.net
Fri Nov 10 16:58:37 EST 2006


Bryan J. Smith wrote:
> From:  Dan Cherry 
>> 32 bit is all I'm running with Xubuntu
>> (and I like it just fine).
> 
> On AMD processors with only 1-2GB, don't bother with 64-bit. And on Intel processors, really don't bother (long story).
> 
> I run it largely so I'm professional exposed to the issues. I also run a handful of apps where the extra features make a sizeable difference. It's more than just the ALU width.
> 
> The "big negative" is that 64-bit is less efficient with memory usage. Not just the 64-bit objects, which is negligible, but the loading of 32-bit libraries alongside their 64-bit ones.
> 
>> the chroot approach is why I left Debian in the first place
> 
> I tried it and gave up myself. I'm too lazy. So if something makes me work, I don't deal with it. Other than a few Firefox details, Fedora Core (especially 6) now makes it less painful.
> 
>> - and I don't see any difference at the user level
>> between 32bit and 64bit 
> 
> Most people won't.  It's really a workstation/server argument until 4GB is required on desktops.
> 
> In fact, the way Microsoft implements Windows x64 - even for Vista - it is _slower_ than 32-bit.
> 
>> the only reason to switch would be for setting up a
>> MythTV box,
> 
> ?  Why  ?

clarification:  switch from Xubuntu to Fedora - not 32bit to 64bit - 
Myth can be set up on either, but I took a practice stab at it a year or 
two ago on Debian, and fought every inch of the way.  I've got a 
reasonable how-to for setting it up on Fedora, and it appears a LOT less 
painful.  (but not essential)
> 
>> and possibly for better compatibility with some 
>> of the hardware I'm using 
> 
> Huh?  Considering peripheral compatibility and drivers are *0* different between PAE36 (32-bit) and PAE52 (64-bit), I don't see why at all?
> 
> You don't need a PAE52 kernel for full compatibility.  PAE36 kernels support 100% everything.
> 
> The only issue is memory mapped I/O performance beyond 2GB with AMD64.  Intel is *0* different (as it can't protect I/O beyond 2GB).
> 
same clarification as above - Xubuntu to Fedora.  I don't have any 
desire to go 64bit yet.

>> Sometimes I waste an awful lot of time trying distros,
>> apps, window managers, whatever,- when 
>> there's nothing wrong with what I'm using.
> 
> Exactomundo!  Stick with what you know and what works for you.  God knows I wouldn't run Linux if I played around with it that much. I do take the time to change a few things on my desktop every now and then, but I  _avoid_ doing so most of the time.  Because I just don't have time.
> 
Agreed, when I switched from Debian to Ubunto, the Debian update cycle 
was causing some headaches, so it was well worth trying a few other 
distros to overcome those problems.  But this time, nothing's wrong, 
other than an over-active curiosity.  And that doesn't really have much 
affect on what runs or doesn't run on my pc's.
> Unless you already run Fedora Core and want to learn/support multi-arch, I would recommend _against_ anyone switching to Fedora Core 6 just for x86-64.  At least not for your main desktop.
> 
> If you really want to "tinker," get a cheap, low-clock Sempron 64 / Athlon 64 for $50, a mainboard for $50 (nForce 4x0/GeForce 61x0) and scrounge up some DDR and disk.  Install Fedora Core 6 x86-64 and use it as a server for remote X desktop solution, not as your primary.
> 
>> unless of course others who have tried FC6 have
>> something to say that I just can't live without ;-)
> 
> I haven't seen Fedora Core 6 offering anything for the typical desktop user over Ubuntu. And I say that as a well-known Red Hat apologist who runs Fedora Core myself. Always have, probably always will.

I think that was the advice I was looking for...   Again, Thanks
> 
> _______________________________________________
> Leaplist mailing list
> Leaplist at leap-cf.org
> http://lists.leap-cf.org/mailman/listinfo/leaplist


-- 
Dan Cherry
dscherry (@) bellsouth.net

Finding a solution to a problem doesn't solve the problem...
Implementing the solution solves the problem.


More information about the Leaplist mailing list