ðòïåëôù 


  áòèé÷ 


Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 

  óôáôøé 


  ðåòóïîáìøîïå 


  ðòïçòáííù 



ðéûéôå
ðéóøíá












     áòèé÷ :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[filmscanners] Re: Digital Darkroom Computer Builders?




"Laurie Solomon" <laurie@advancenet.net> writes:
>>>>>>>>>>>
>A 32-bit address space can only span 4 gigabytes.  I can't go past 4
>gigabytes with a 32-bit architecture.

While that makes clearer what you are saying in one sense; it still is
unclear if you are saying that it is impossible to do with 32-bit
architecture or merely difficult or inefficient to do.  There is a
difference.  One is a brick wall limitation while the other is a limitation
by choice or preference.
<<<<<<<<<<<<

The 32-bit architecture of the CPU in user mode has nothing to do whatsoever
with how much physical memory the OS can handle. That is determined by the
MMU and the width of the address bus the MMU provides. Under Win NT, Win
2000, and Win XP user processes are limited to a 2GB (and in one version
3GB) of address space. Win NT, Win 2000, and Win XP normally are limited to
4GB of physical memory, but the "Advanced Server" and "Datacenter" versions
are limited to 8GB and 64GB of physical memory respectively. (Advanced
Server can be booted in a mode that gives the user 3GB of memory. But (the
last time I looked) Advanced Server is expensive.)

The bottom line is that 2.5GB is the largest amount of physical memory it
makes sense to have on a normal XP machine, and that gives Photoshop 2GB to
play with.

While it turns out that 2GB on a somewhat over 2GB machine is the most we're
going to get for image editing on 32-bit 'Doze, Anthony's claim that
handling more memory than an individual instruction can access is both
innefficient and difficult is wrong on both counts. Accessing the whole of
the address space from every instruction is hideously inefficient. Most
machines provide modes where a base register plus a short offset field in
the instruction is used. This is much more efficient than including the
whole address in every instruction. At which point, the size of the base
register is the only limit on program address space. Furthermore, if you are
programming in any language other than assembler, the compiler (or
interpreter for you Scheme and Java fans) will take care of all of that
transparently. (Although it helps to understand what the compiler is doing
so as to allow it find efficient ways of doing things.)

>>>>>>>>>>>
But I have also been suggesting that first this discussion is
not about you and your personal practical situation and second you have a
tendency to project your personal limitations on everyone else as a
universal rather than as a special case or an individual case.
<<<<<<<<<<<<<<

Really.

David J. Littleboy
davidjl@gol.com
Tokyo, Japan



----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body



 




Copyright © Lexa Software, 1996-2009.