****Re: ****Re: ****Re: ****What's up with 64 bit Linux

[Craig White]

On Fri, 2007-11-23 at 00:01 -0700, Chris Gehlker wrote:
> On Nov 22, 2007, at 10:32 PM, Craig White wrote:
> 
> > On Thu, 2007-11-22 at 18:46 -0700, Chris Gehlker wrote:
> >> On Nov 22, 2007, at 5:49 PM, Jon M. Hanson wrote:
> >
> >> I readily admit that what I can easily find on the web corresponds to
> >> the "Compatibility mode can only be set at boot time so it's not
> >> something you can switch back and forth while the system is running"
> >> model that you lay out above and that model is perfectly consistent
> >> with what Craig reports. Something is missing though. These machines
> >> can run either 32 or 64-bit Windows but only 64-bit Linux. It doesn't
> >> make sense.
> > ----
> > but that isn't what I said at all.
> >
> > It's not that 32 bit Linux can't run on these systems...it's that the
> > code within the i386 kernel and modules doesn't work properly and I am
> > reliant upon open source developers, mainly kernel developers to work
> > through the issues presented by this specific hardware.
> >
> > The fact is that even on the 64 bit versions, there has been  
> > regression
> > within Fedora 7 which used a 2.6.21 kernel when released and the last
> > update used a 2.6.23 kernel and now the systems no longer poweroff  
> > when
> > instructed to shut down (not a big deal but a minor nuisance).
> >
> > The 2.6.23 kernel used on Fedora 8 is even worse for this specific
> > hardware. That hardware is not available with Windows XP
> > pre-installed...only Vista. Obviously there is some support for this
> > hardware within Windows Vista that isn't incorporated into 2.6.23
> > kernels (yet at least, perhaps in a test kernel that I am not aware  
> > of).
> 
> 
> OK, So your experience is that i386 runs but not as well as i86_64.
----
today...on Fedora 8 and a Dell Optiplex 320 (note the very strained
conditions), there is little to no difference installing i386 or x86_64
bit version. I hope tomorrow will change that.

I always have to keep in mind that Fedora is somewhat edge software -
it's not reasonable to expect 2.6.23 kernel stuff to be as stable as say
2.6.21 or earlier.
----
>   I  
> notice from that link you cited that the Fedora folks recommend the  
> opposite approach.
----
I'm not sure what you mean by 'opposite approach' - their
recommendations are based solely upon the processor in the hardware you
are using
----
> > As for 64 bit Windows...it's there but few are using it because it  
> > is so
> > incredibly buggy that you have to be a martyr to use it so I'm told.
> >
> > It's easy enough to boot a 64 bit Linux and run 32 bit versions of
> > programs provided that you have the library support to do so. Many
> > choose to do that for things like Firefox so they can use 32 bit only
> > versions of plugins such as Adobe's Flash and Acrobat Reader.
> 
> OK then Jon's information about having to pick a mode at boot time is  
> wrong. But the real problem here is that while the advantages of  
> running 64-bit apps on a 32-bit OS are clear, there aren't any  
> advantages to doing it the other way around.
----
I gathered that the point Jon was trying to make was that the kernel
code loaded at boot signals the 64 bit processor to either emulate 32
bit operations or 64 bit operations and that could not change until
reboot.

As for what may be advantages of 64 bit applications on a 32 bit OS...I
am not aware that such a thing is possible but this is beyond my
knowledge base
----
> > I just wanted to point out that you are drawing conclusions that I  
> > don't
> > believe are supported by my own experiences.
> >
> > Looking at this matrix that is part of Fedora 8 installation
> > documentation,
> > http://docs.fedoraproject.org/install-guide/f8/en_US/sn-which- 
> > arch.html
> > it's pretty clear that they are recommending installation of the 86_64
> > version for all Intel Core 2 Duo, Centrino Core 2 Duo, Xeon and AMD 64
> > processors.
> 
> Yes. But they don't say why. When I took an operating systems class,  
> not that long ago the goal was clearly to avoid 64-bit code whenever  
> possible because it page faults more. Of course now that I look again  
> at the referenced page, it doesn't say that x86_64 is actually 64-bit  
> code.I could be mostly 32_bit code optimized for 64-bit processors.
----
I almost suspect that this classroom knowledge tidbit probably was more
specific than an overarching rule - i.e., this might be a Windows issue
as I understand that there are still many of these issues that continue
to plague the 64 bit Windows OS.

Again, we are over my head here and I am not comfortable spouting things
that I know so little about but I am under the general impression that
x86_64 code is actually 64 bit code.

Craig