Gus wrote:
> This is what is known as PAE (Physical Address Extension). It works
> much the same way as the original 8088 did with segments and offset
> registers, allowing you to access more RAM than the baseline register
> size.

I never really understood assembly on x86 and this is one of the
reasons why. How on earth did this processor ever get so popular when
it was such an ugly PITA to code for? I still have my first x86
assembly book on my bookshelf. I made it about halfway through before
I ran into stuff that I just couldn't debug because I couldn't
understand it. I would probably do much better these days but I see no
point in wasting my time on such things as x86 real mode. This was
around 1989 or so.

For anyone who wants to do x86 assembly these days I recommend
"Programming From The Ground Up" by Jonathan Bartlett. I have the hard
copy but you can legally download a pdf from here:

http://mirror.sourceshare.org/savannah/pgubook/

This book does assembly programming in Linux and therefore everything
happens in protected mode and your program is presented with a linear
view of memory. The book examples are quite simple but well
explained. It also explains how to link your assembly code with other
higher level code such as C or Python. So you can write the
performance intensive parts in assembly and UI and other stuff in a
high level language.

So, why couldn't they make the registers bigger instead of adding more
registers to index segments?

On Mon, Mar 09, 2009 at 09:52:27AM -0700, Barry Gershenfeld spake thusly:
> can go above 1 megabyte!" Then, the inevitable one-upmanship: "Hey, we 
> can store that bit in the keyboard controller!"

The extra bit was stored in the keyboard controller? Knowing x86 I
can't tell if this is a joke or not.

> [3] Maybe none of this matters anymore, now that memory management is  
> virtualized

I hear the instruction decoding and MMU on an x86 machine take up huge
amounts of power and chip real-estate. Will this ever become such a
problem (due to increasing numbers of cores and power requirements)
that it becomes a significant disadvantage to x86?

> [4] Who thought up this lousy 64-bit limitation, anyway? :)

18 exabytes should be enough for anyone! But I bet we'll have to stop
saying stuff like this eventually. I hear that 2^128th is enough to
account for every atom in the universe or something like that.

-- 
Tracy Reed
http://tracyreed.org

Attachment: pgp7gNpitVSPz.pgp
Description: PGP signature

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-newbie

Reply via email to