PDA

View Full Version : Inquirer article on Linus on Intel x86-64 revolution



Kooldino
24th February 2004, 10:16
heh...

http://www.theinquirer.net/?article=14301

Sorry if this is a repost, i searched first.

So is Intel actually using the same exact standard and instruction set that AMD came up with already?

Chrono_Wanderer
25th February 2004, 18:50
yes, i'd notice they never give credit to AMD for x86-64.

and yes, according to Intel, its completely competible with the x86-64 standard.

but, technically, IA-64 IS a lot better than x86-64, and I think that's why IA-64 tooks so long to R&D. x86 is just dead old, from some 30 years ago. Amazing that 30 years later we are still using it.

All these backward competibility crap, from 8bit all the way up to 64bit.

Its not like we need Windows 1.0 anymore.

The problem why we can't drop x86 is because the competibility needs to be continuous....

spadnos
26th February 2004, 18:49
There was a line in the full series of emails (with Linus and others) that said something like "It's the same, except for the little details that are always different between processors".

So, t seems as though it's the same instructions, with possibly different results (like flags being set differently or the execution speed being different or something)

- Steve

Belwarrior
26th February 2004, 20:23
Well... i would be very shocked if Intel is going to let people know it is making an AMD clone...:p:D

Kooldino
27th February 2004, 14:17
IIRC, AMD was just gonna use this line of chips as a transition chip, and then drop legacy in their next line of chips once the software is all 64 bit.

But I could be wrong.

rugger
27th February 2004, 21:20
Kooldino

Doubtful because the backwards compatibility stuff is really just a subset of the processors 64bit abilities. Therefore, the transisitor cost of leaving legacy in the chip is cheap enough that it would be stupid for AMD to remove it and alienate its users.

rugger
27th February 2004, 21:22
So basicly:

X86 isn't going away, no matter how much we hate it, nor how ever much we want it to :D

*evil cackle*

SpiralDragon
27th February 2004, 23:32
too bad.... but its about time we had a REAL major revolution in the PC industry...... i for one would love to see the flopy compleetlty die..... it can be don... but the likes of intel, amd and MS just want to keep on taking our every last peny ... :( ...... IA64 is superior... but if only they'd mainstream it alot quiker

rugger
28th February 2004, 00:47
Originally posted by SpiralDragon
too bad.... but its about time we had a REAL major revolution in the PC industry...... i for one would love to see the flopy compleetlty die..... it can be don... but the likes of intel, amd and MS just want to keep on taking our every last peny ... :( ...... IA64 is superior... but if only they'd mainstream it alot quiker

What for, I really think people don't want a revolution. Its expensive and unsafe road to travel.

It seems perfectly clear that it is just real hard to significantly beat the performance of the x86 architecture on the desktop. X86 instuction emulation seems to be so effective that its unlikely that we will ever need to replace it. X86-64 support from intel is likely to ensure that x86 is around for a LONG LONG time now.

For what it is worth, the floppy drive is more or less dead. Our floppy drives sit around doing nothing for so long they die of dust contamination rather then actual use. CD-R and the internet have made the floppy drive more or less completely obsolete. Removing the floppy drive from new computers would still only save about $30 or so. It simply isn't complex or expensive enough to hold computers back.

As for intel, amd and microsoft wanting our money, of course they want our money, thats why they won't abandon old interfaces that are cheap to include yet ensure people keep buying their products.

az
28th February 2004, 04:31
We don't need ever increasing processing power anyway, we need better software (more efficiently coded, but in the first place with a REALLY good user interface. They should let game designers work on OS UIs.).

AZ

rugger
28th February 2004, 04:47
Originally posted by az
We don't need ever increasing processing power anyway, we need better software (more efficiently coded, but in the first place with a REALLY good user interface. They should let game designers work on OS UIs.).

AZ

NO ... aggghhhhh ... most game interfaces are painfully horrible :/ I would never imagine forcing such UI design onto people who simply want to use their computers.

az
28th February 2004, 05:21
OK, GOOD game designers ;)

Or me ;)

AZ

leech
29th February 2004, 07:26
:D When I first saw the Enlightenment Window Manager, I thought that it LOOKED like a video game's interface. That was back in the DR0.13 days, and that's what made me try out linux in the first place. Because it was just COOL looking.

I agree with Az in the fact that they need to have REALLY good code now. Many times over I have complained on MURC about why the code just keeps getting bigger and bigger. Programmers these days are too lazy to optimize their code as much as they had to back in the 8/16bit era. It's better for them to just make a game or app that requires more horsepower so that people are forced to upgrade often.

A great example of this is Frontier: Elite 2. That game for the Atari ST fit on one 720kb floppy disk (the actual game, which had a decompression routine built-in, was only about 360kb.) That game was HUGE. Granted it didn't have a million textures all in 24bit color, but I think you get my point. If people would just optimize their code for the hardware better, life on the PC would be much nicer.

Leech

rugger
29th February 2004, 20:50
Originally posted by leech
:D When I first saw the Enlightenment Window Manager, I thought that it LOOKED like a video game's interface. That was back in the DR0.13 days, and that's what made me try out linux in the first place. Because it was just COOL looking.

I agree with Az in the fact that they need to have REALLY good code now. Many times over I have complained on MURC about why the code just keeps getting bigger and bigger. Programmers these days are too lazy to optimize their code as much as they had to back in the 8/16bit era. It's better for them to just make a game or app that requires more horsepower so that people are forced to upgrade often.

A great example of this is Frontier: Elite 2. That game for the Atari ST fit on one 720kb floppy disk (the actual game, which had a decompression routine built-in, was only about 360kb.) That game was HUGE. Granted it didn't have a million textures all in 24bit color, but I think you get my point. If people would just optimize their code for the hardware better, life on the PC would be much nicer.

Leech

I'm gonna call bullcrap on this argument too. You really need to stop living in the 80's and get a grip on reality. That reality is:

OPTIMISATION IS EXPENSIVE
OPTIMISATION INTRODUCES BUGS AND IS DIFFICULT TO DEBUG
OPTIMISATION TAKES A LOT OF TIME

So, for any software house writing software, they have the option of releasing eariler, but requiring somewhat more powerful machine to run, or they have the option of releasing later, with more bugs, at a higher cost, but requiring a somewhat less powerful machine to run. Not only that, but in the time the software house spent optimising, more powerful hardware is being released making any optimisations they do less useful.

As a customer, would you prefer a more optimised program over one with fewer bugs, at a cheaper cost, and delivered earlier.

Also ask yourself if it would be better to spend a little bit of money upgrading hardware rather then paying a lot more for optimised software.

Note: my rant does not excuse excessively bloated software like ICQ, which no rant could ever excuse :D

Jon P. Inghram
29th February 2004, 21:16
As a customer, would you prefer a more optimised program over one with fewer bugs, at a cheaper cost, and delivered earlier.

Judging by the popularity of Windows, the answer is neither. ;)