conspiracy theory - Page 3

I wouldn't necessarily equate the jump in AMD's 64 bit performance in Linux to what it will be in Windows, or what you can expect from Intel for that matter. First reason is that Linux is typicall a leaner OS (though Linux is stating to suffer from bloat) and doesn't require as much resources, so it can have better performance. Not only that the whole kernel can be recompiled to suit the environement that Linux is running in. That drops some dead weight aside from other tweaks you can do to the OS.

Also keep in mind that AMD has an integrated memory controller which can access memory a little faster. The integrated controller has much lower latency than the typical PC we use now.

As it stands now, x86-64 (at least how AMD has implemented it) had 3 different modes: essentially an enhanced which runs straight 64 bit, a "hybrid" mode which allows a mix of 32 and 64 bit code to run simultaneously, and a "legacy" mode that can run in pure 32, like most Hammers running Windows right now. The attractive thing is that 32 bit can be run natively without a performance hit and will run even better in straight 64 bit mode.

I guess x86-64 is good and bad in ways. It's good that we can run our stuff without a performance hit, but it just means that it'll be a LONG time before we can get the really good stuff like Itanium class processors on the desktop (288 general purpose registers vs 16 with x86-64 ( ).

Super

Not waiting for Intel.

Waiting for the market.

That, and they're still working on Longhorn

and who needs 16GB of ram now, anyways...

Intel pretty much is the market. AMD has roughly 15-20% of the market, and with a relative handful of their marketshare cranking out Hammers (probably around 10% of their total at best a quarter now), that would put Hammer penetrating at around 2-3%.

Not too profitable for MS at this point. Intel can and will push that figure upward to make it worth MS's while.

Longhorn's still about 2 years out at least. End of 2005 is very optimistic. mid to late 2006 is probably a bit more realistic as Longhorn is a major overhaul.

That's too far out and MS wants to make money in the meantime, so XP for 64 bits is their best chance to do that in the interim.

Super

a64s arnt bad, but DO NOT CONSIDER 64-BIT SUPPORT A REASON FOR BUYING IT. its like buying a car based on how well it could stand an atomic blast. no matter what your freakin dead. and likewise by the time there is a any reason to switch the 64-bit os, your computer will suck so bad the performance increases won't even apply to you.

dunno about that man... using 64bit Linux gives you a 20% boost now... that's worth buying

also, MS are delaying xp64 because they have still not released xp sp2... they need to get that out the door, deal with the problems, then release xp64

Intel is not the market.

Consumers are the market.

Profit does not come from increased production, but rather directly from buyer demand, and right now it's more of a problem in advertising than manufacturing.

There simply isn't a need right now for 64bit technology in the PC market.

Anyway..

IIRC, Microsoft plans on a revision of WinXP before the Longhorn launch.

It will be to WinXP what 98se was to the original 98.

with what software exactly?

using 64 bit linux to do what? run a server?

With all the billions MS has, you don't think MS would have at least two separate teams working on those?

Ok, how about Intel corners the market?

Profit can come from increased production. Hammer prices are high right now for two reasons IMO:

AMD prices their chips as much as Intel. They may be worth the price, but AMD consumers typically are used to paying much less than Intel prices for similar performance. There's a reason why cheap XP's are still AMD's bread and butter. Add to that the fact that AMD thinks it has something on Intel that it can charge extra for. Unfortunately, there's nothing besides Linux to exploit it and that won't appeal to the mainstream market.

Secondly, you linked it to supply and demand. Supply is low, whether purposefully or not, so prices are going to be higher. That will keep the prices up and demand lower from AMD's core demographic. Drop the price some and I bet more people would buy them. You'd think it'd be better to sell more chips at somewhat less money then sell few chips and get more for them. Selling more chips would give them more of an advantage in the marketplace, especially when Intel's vulnerable.

Remember, AMD said the market would be flooded with Hammers by now, and the market's nowhere close to that.

Also keep in mind that there really isn't THAT good of a platform out for A64 yet. That will affect demand as well.

You're right about advertising, AMD refuses to invest in itself. I've only ever seen one AMD commercial on TV my whole life, whereas I hear the Intel chime several times daily. I see the occasional ad in the paper or on a website, but Intel's still out advertising them.

I think XP for 64 bit systems will be a bit more than 98 to 98SE. You're adding a whole new codebase to it, and not everyone will feel compelled to buy it if they don't have the hardware to support it, whereas 98SE was more a drop in replacement.

Super

i read somewhere that the 64 bits really helps for photo work. i bet it would help for video design as well. I really don't think there will be a 3 year delay either for this new operating system. Also i doubt that intel has 80% of the processor market. there are other companies out there, one being ibm that creates computer processors.

I think photoshop and the like will benefit, no doubt.

I probably should have said desktop market. Intel does well in the server market with the Xeons. They're small in the big iron where IBM, HP and Sun dominate, but they're starting to crack the market a bit better.

Now that I'm thinking of this, I'm curious. I'll have to check the bitness of my Sun workstation at work.

Notebook technology tends to last a lot longer than PC technology. You often see real old NB's becoming e-mail checkers/word processers, ect. I would be tottaly happy with being able to run a 64bit based OS 2 years from now while sitting in bed checking E-mail/web browsing. I think that could maybe (and i mean maybe) apply to PC technology, but as far as Notebooks, i plan on keeping this for several years and doing all my initial testing on 64bit beta's on this machine. Time will tell how the new software will perform, obviously in theory it must be better, otherwise they wouldn't be designing it now would they

you have a point, the portability of a laptop is not to be underestimated. You could keep a 64-bit amd lappy hanging around for over a decade after it can't handle games just for surfin and email in bed or whatnot.

i think that the AMD 64 is a few dollars more expensive right now actually.

its not about what technology is best right now. its about timing. switching to newer products too fast would make older ones more obsolete.

there is a lot more money to be made with XP. switching to XP64 would take quite a bit of market share from its own older product.

lets say, 1/2 the people in the world own XP right now. bringing out XP64, only 1/2 of those people upgrade. now waiting for more to get XP, means more upgrading, means more money.

unfortunately, Microsoft doesn't have many competitors, so they can do this timing strategy.

with computer parts and what not, competitors have to bring out better products before others do, so they can't do this timing (controlling of the market).

yeah... right

I think it works like this.

Lets make Longhorn -- OK, good idea.
Lets make XP64 -- OK, good idea.
Lets make XPSP2 -- OK, good idea.

[ time passes ]

Oh sh!t, its near XPSP2 release time... if we add more ppl to the project, and confuse those already in it, it should be done faster [ nix Longhorn for 3-6 months].

[ time passes ]

The moral of "The mythical man month" applies at this point... but managers being managers decide more ppl are needed! [ nix XP64 for 3-6 months].

[ time passes ]

XPSP2 is released, with lots of bugs due to mismanagement. [ XP64 goes back into production ]

[time passes]

Some bugs are fixed in XPSP2. [ Longhorn goes back into production ]. Everyone is happy, lots of extra money has been spent.
Managers call a lessons-learned meeting, everyone resolves to read "The mythical man month", but no one does... repeat process with next piece of S/W.

I think Microsoft knows what it's doing.